A new licensing deal between Pure Storage and an as-of-yet unnamed hyperscaler could precipitate a transformation in the enterprise storage market, according to Pure Storage CTO Rob Lee.

Flash storage has long been hailed as the successor to its spinning disk counterparts, yet thus far HDDs have clung on in the enterprise market. This could all change, according to Lee.

In November 2024, Pure Storage announced it had reached its first flash storage deal with a hyperscaler.

The hyperscaler in question is set to be revealed “soon enough”, according to Lee, who added this is the first of what the company hopes is a number of similar agreements with the other major industry players.

“We recently announced our first, hopefully, of many design wins with a top four hyperscaler who’s chosen our technology to standardize all of their storage, from their cheapest disk all the way to eventually their highest performance SSDs” he stated.

During an earnings call following the announcement, which caused Pure Storage’s share price to soar, Charlie Giancarlo, CEO at Pure Storage, said the deal could mark a turning point in the market, describing it as “the vanguard for flash storage providing all online storage in major hyperscale environments in the future.”

This could significantly shake up the enterprise storage market, Lee said, noting that the market for disk storage is incredibly concentrated on both the supply and demand side, hinting that one big move from a hyperscaler could have a major effect on the industry as a whole.

“There’s two suppliers of disks, and 60 to 70% is purchased by the top five hyperscalers. Just let that sink in for a second. So when you have a market that is that concentrated, if one of those buyers flips and stops buying disks and shifts over to flash, you can see what’s going to happen,” he explained.

“I think that’s going to be the biggest catalyst, the biggest nail in the coffin finishing the transition from disk to flash.”

Flash storage could be an unsung hero for decarbonizing hyperscalers

Lee’s comments came during a press tour of Pure Storage’s global R&D center in Prague this week. During a Q&A session, he told ITPro and assembled journalists the factors that brought the hyperscaler to the table included better reliability and performance.

The primary reason, however, focused firmly on the energy savings flash can bring compared to traditional storage methods.

Hyperscaler capital expenditure has been growing as they race to dominate the burgeoning AI market, but this has presented major challenges in terms of energy security, Lee noted.

Nearly all of the top hyperscalers now have projects to source clean energy to accommodate for soaring demand, with Google, Microsoft, AWS, and Oracle all signaling their approval for nuclear energy in particular.

With this in mind, Lee said Pure Storage has been working with hyperscalers over the last two years to both demonstrate and emphasize the benefits of flash storage in data center operations, adding that a single digit decrease in energy consumption could be enough to drive a wholesale transition to flash.

“These guys operate at such a scale that saving 1 or 2% in power overall is a huge needle-mover for them. By uptaking our technology or storage layer, they’re actually going to save more energy than building nuclear power plants.”

HPE’s Primera Tier-0 storage arrays are designed for enterprises that simply cannot tolerate any downtime. Whereas many competing products use an active/active controller design, the Primera employs an all-active mesh architecture that has no passive components - allowing each volume to be active on all controllers all of the time.

Central to the controller nodes are HPE’s application-specific integrated circuits (ASICs) which are geared up for NVMe performance. A key feature is the CPUs and ASICs in each node are all interconnected so unlike arrays that have to mirror cache contents across controllers, all node cache memory in a Primera array is presented as a global, fault-tolerant pool. The Primera’s OS also comes into play as its services are implemented independently, which means the array can be upgraded or updated without requiring a reboot.

HPE is so confident in the Primera’s architecture that its Proactive Care support package includes a 100% data availability guarantee. There are some provisos such as ensuring the array is kept fully updated but if an outage is traced directly to Primera, HPE will provide a credit of up to 20% of the original purchase price towards future Primera purchases or upgrades.

HPE Primera A630 review: Hardware line-up

In this hands-on review, we’re testing HPE’s entry-level Primera A630 all-flash array. This 2U model employs dual, hot-plug controllers each with 128GB of cache memory, a network management port and dual embedded 10GbE SFP+ ports specifically for HPE’s remote copy over IP (RCIP) replication service.

Port options look good: the controllers have two PCI-E expansion slots and we were supplied with a quad-port 16/32Gbps fibre channel (FC) HBA in each one. A maximum of 16 ports can be installed and HPE also offers quad-port 10/25GbE and 10Gbase-T HBAs.

The array presents 24 hot-swap SSF bays at the front and our system came with a full house of 1.92TB SAS3 SSDs. The last 8 bays are dual-function and can support NVMe SSDs while the SAS3 expansion ports on each controller allow up to 5 enclosures to be added for a total of 144 SSDs.

HPE Primera A630 review: Deployment

Initial deployment is swift: we used HPE’s discovery tool to find the array and after logging in to the Primera OS web console, followed the quick start wizard and had it ready for action in 20 minutes. You don’t need to worry about choosing a RAID type as the Primera automatically divides each physical drive (PD) into 1GB ‘chunklets’ and implements RAID6 across them all.

With our 1.92TB SAS3 SSDs costing nearly seven grand a pop, this has distinct advantages; none are sitting idle and load balancing means their life span is prolonged. HPE has also simplified the process of joining hosts with virtual volumes (VVs) as you place multiple hosts in one set, add VVs to another and bring them all together with an application set.

Using two Xeon Scalable servers running Windows Server 2019, we cabled them to the array over dual 16Gbps FC connections to each controller and could see that our FC MPIO paths were all active and optimised. We then created a host set for both servers, selected their FC paths and added them all to an application set which already had two 5TB VVs ready and waiting.

HPE Primera A630 review: Management

Management options are extensive. Along with the Primera OS console, we added the A630 to the lab’s Hyper-V virtualized HPE SSMC (StoreServ management console) instance. We were presented with an overview of all HPE storage systems and could drill down to the A630 and check on its system status, performance, volumes, capacities, host bandwidth and more.

If you want to use RCIP, you’ll need SSMC as this can’t be configured from the Primera OS console. Furthermore, although you can use the Primera OS to assign snapshot schedules when creating application sets, they can only be managed from SSMC which is also required for creating virtual copy snapshots.

The Primera was easily integrated into the lab’s HPE OneView 6.3 virtualized instance and those with an HPE Greenlake account can remotely manage it from the Data Service Cloud Console. HPE’s InfoSight is the star of the show though, as it applies AI and ML to historical performance data collections to predict future system loads and provide capacity forecasts.

The only problem was, we couldn’t get it to work. Even though the A630 clocked our HPE Passport account organisation, it resolutely refused to register with it and while this resulted in plenty of head-scratching at HPE, no resolution was forthcoming. To be fair, our review system had seen plenty of previous action in customer sites so its configuration may have been messed about with - but it was disappointing that HPE was unable to fix this and reveal all of InfoSight’s glories to us.

HPE Primera A630 review: FC performance

Starting with one server, Iometer reported speeds of 40.4Gbits/sec and 42.4Gbits/sec for sequential reads and writes with random operations returning nearly the same numbers. Throughput was also good, and for sequential operations Iometer recorded 648,300 IOPS and 165,500 IOPS, with random operations returning 653,500 IOPS and 177,600 IOPS.

Adding our second server to the mix returned cumulative sequential reads and writes of 80.8Gbits/sec and 48Gbits/sec while random numbers came to 80.8Gbits/sec and 47.9Gbits/sec. It was the same story for throughput, too; sequential and random reads and writes delivered 802,600 IOPS and 247,500 IOPS and 799,800 IOPS and 273,700 IOPS respectively.

In our throughout tests, we also noted that latency was impressively low, ranging between 0.4ms and 1.5ms. We did find that running our tests on VVs with HPE’s advanced data reduction enabled resulted in reduced write speeds, but this didn’t have any impact on latency.

HPE Primera A630 review: Verdict

HPE’s Primera A630 and its innovative all-active architecture makes it a great choice for delivering Tier-0 storage to business-critical apps. We found it remarkably easy to deploy and although HPE failed to get it talking to our InfoSight account, general management features are very impressive. FC performance is good with the array exhibiting extremely low latency, while HPE’s 100% data availability guarantee will be very reassuring for organisations where downtime is a dirty word.

HPE Primera A630 specifications (as reviewed)

HPE’s MSA storage arrays have always had a firm focus on SMBs and the latest Gen6 2060/2062 models aim to deliver more power for even less cash. Their uprated ASIC and larger cache memory claim up to 45% IOPS and 75% sequential throughput increases over the previous Gen5 2050/2052 arrays, and the latest firmware brings in a number of valuable new features.

As with previous models, the new MSAs come with dual controllers running in active/active mode and their cache capacities go from 8GB to 12GB per controller. There’s no need for batteries to protect the cache contents in the event of a power failure either, as each controller employs a fast charging super-capacitor and CompactFlash card.

A significant change, however, is the new controllers no longer offer converged host ports. Whereas the 2050/2052 arrays allowed you to install the desired transceivers to set their personality, the 2060/2062 versions must be ordered with either 8/16Gbps FC or 10GbE iSCSI host ports – HPE also offers a 12Gbps SAS3 direct-attach option.

The arrays are available with 12 LFF or 24 SFF hot-swap drive bays supporting SAS3 and Midline SAS (ML-SAS) HDDs and SAS3 SSDs. Expansion potential also sees big improvements; the port on each controller goes from SAS2 to SAS3 and supports up to nine disk shelves for a total of 240 drives, bringing a disk density increase of 25% over the Gen5 arrays.

HPE MSA 2060 Storage review: Storage features

On review is the MSA 2060 Storage FC model, which comes with four 8/16Gbps FC host ports per controller. The software offers plenty of standard storage features including thin provisioning, snapshots, volume copies and SSD read caching.

Storage pool RAID options are extensive, with support for striped SSD read caches plus standard mirrors, RAID10, 5 and 6 arrays. HPE also introduces its own MSA-DP+ virtual disk groups which support from 12 to 120 disks and are designed to improve performance and provide faster rebuild times.

HPE’s data tiering supports a performance tier of SSDs, a standard tier of enterprise SAS drives and an archive tier of ML-SAS drives. The smart bit is all page moves between the tiers are fully automated, allowing the array to respond quickly to changes in I/O patterns.

The standard and archive tiers are included by default on the MSA 2060, but the performance tier is optional and requires HPE’s MSA Advanced Data Services Suite (ADSS) license applied. Included with the 2062, it’s worth having as it also increases the number of volume snapshots from 64 to 512 and activates the remote snap service for asynchronous replication to remote MSA arrays.

HPE MSA 2060 Storage review: Deployment

When we reviewed the MSA 2052 we thought installation was easy enough, but HPE has simplified it even further. On first contact with the new SMU v4 web interface, it kicked off with a wizard that guided us through system setup, securing management access, creating storage pools, provisioning virtual volumes and setting data port host assignments.

The controllers support one storage pool apiece which can be up to 1PB in capacity. Each contains multiple disk groups which are created by selecting disk types, RAID levels and the controller you want to assign them to.

Our review system had the ADSS license applied so when we assigned a dual SSD mirror to a pool, it was automatically placed in a performance tier while our SAS3 HDD arrays went into a standard tier. If you don’t have the ADSS license, you can still use SSDs to create a read cache and assign it to a pool.

Creating virtual volumes is a cinch. We simply chose the controller to determine the pool they would go in, provided a name, entered a size in MB, GB or TB and assigned them to specific host data ports. Snapshots are created by selecting the data protection option for a volume and running them on-demand, or applying a schedule where they can be repeated as often as every minute.

HPE MSA 2060 Storage review: Performance

For performance testing, we used Dell and HPE dual-socket Xeon Scalable servers running Windows Server 2019 and equipped with quad-port 16Gbps FC ATTO Celerity HBAs. Over on the MSA 2060, we created two pools, each with a dual SSD mirrored performance tier and a RAID5 standard tier made up of five 10K SAS3 HDDs.

When creating MPIO links, bear in mind that the array uses an asymmetric active/active mode. If you connect a host to each controller, you don’t get a doubling in performance, as the link to the controller that doesn’t own the pool is designated as non-optimised and reserved for automatic failover.

With one server connected across both controllers and assigned a 1TB volume from Pool A, we started Iometer and watched all activity shift over to the SSD performance tier in around 5 minutes. Sequential read and write performance were both measured at a steady 12.3Gbits/sec while random rates settled at an equally speedy 12Gbits/sec and 7.4Gbits/sec.

Assigning a 1TB volume from Pool B to our second server saw cumulative sequential read and write speeds for both servers ramp up to 24.6Gbits/sec with random rates increasing to 19.1Gbits/sec and 14.8Gbits/sec. Cumulative IOPS were equally good - Iometer reported 640,000 and 178,500 IOPS for sequential reads and writes, with random operations returning 398,500 and 111,000 IOPS.

HPE MSA 2060 Storage review: Verdict

Although we tested HPE’s MSA 2052 Storage over 10GbE iSCSI, we can still see that the Gen6 MSA 2060 FC model and its uprated hardware and larger cache delivers significant performance improvements across the board. It retains all the features that make it appealing to SMBs such as ease of deployment and automated data tiering, it offers an even greater expansion potential and the comparatively low starting price makes it extremely good value.

HPE MSA 2060 Storage specifications

HPE's MSA arrays are very popular with SMBs as they offer a great range of storage virtualisation features at unbeatable prices. The latest Gen5 MSA 2050/2052 appliances look like even better value, offering a doubling in performance for the same price.

Along with an updated Broadwell CPU, the new controllers are endowed with twice as much cache memory. Each one has 8GB of cache and employs a super-capacitor and CompactFlash card to protect its contents in the event of a power failure.

Connection options abound as the converged controllers on the SAN models each have four SFP+ ports grouped into two pairs. Each group supports 8/16Gbps Fibre Channel or 10GbE iSCSI. Simply install the desired transceivers to set their personality - remembering that the two controllers must match.

HPE also offers a 4-port 12Gbps SAS3 controller model for direct attached storage. You can have 12 LFF or 24 SFF hot-swap drive bays and all models support SAS3, Midline SAS and SSDs while the 2050 can also handle self-encrypting drives.

Storage virtualization features

Standard software features across both the 2050 and 2052 models include thin provisioning, snapshots, volume copies and SSD read caching. The stand-out feature is HPE's automated data tiering, which employs a performance tier comprising SSDs that stores hot data and delivers a big performance boost.

This is optional on the 2050 while the 2052 includes HPE's MSA Advanced Data Services Suite, which increases the number of snapshots and activates the performance tier. Standard and archive tiers are included with all models, where data is initially held on SAS drives and moved to lower-cost Midline SAS drives as it gets cold.

The appliances analyse data usage in real-time with 4MB page moves occurring every 5-10 seconds. This allows them to respond quickly to changes in I/O patterns making them a far better choice for changing or unpredictable workloads than many competing SMB arrays that only use daily I/O batch processing.

Deployment

You don't need to be a storage guru to deploy the MSA 2052, as its slick SMU v3 web console kicks off a nine-step quick-start wizard to help configure management access, system settings and data port assignments. Our review system was supplied with a full house of 10GbE SFP+ iSCSI transceivers so we just had to assign static IP addresses to each one.

Storage is just as easy to provision as the appliance supports two pools assigned to each controller. Pools contain multiple disk groups, which are created by selecting disk types or SSDs, RAID levels and the controller they should be assigned to.

There's nothing else to do; the appliance automatically places SSDs in the performance tier, SAS drives in the standard tier and Midline SAS drives in the archive tier. If a pool isn't using a performance tier, you can create SSDs read caches if you wish.

To create virtual volumes, we simply chose a name and a size, the number of volumes we wanted and assigned it to a controller to determine which pool it would go in. You can leave the appliance to manage all data movement or select a tier affinity to force it to use the performance or archive tiers.

Snapshots and recovery

Previously, the SMU console only supported manual snapshots of virtual volumes but this has been updated to include a scheduler. You can still create on-demand snapshots but the console now allows you to schedule them to repeat as often as every minute, add a daily time window and decide to retain up to sixteen snapshots.

Snapshots are very versatile as they don't require a dedicated pool and you can take child snapshots (snapshots of snapshots). You can reset snapshots so they contain the latest data and map them as new volumes whenever you want.

Recovering data from a snapshot is a swift three-click manoeuver. We tested this by deleting some data from a virtual volume, selected the Rollback option, chose a snapshot, unmounted the volume from connected hosts and had the data restored in less than 10 seconds.

Performance

For testing, we used two Windows Server 2016 hosts both equipped with dual-port 10GbE SFP+ adapters. One was connected to two ports on Controller A and the other to Controller B to allow us to create load-balanced MPIO links.

HPE supplied our test system with four SSDs, eight SAS drives and twelve Midline SAS drives. We created two pools each with three tiers and mapped dedicated virtual volumes to each host.

IP SAN performance is excellent, with Iometer running on both servers returning cumulative sequential read and write speeds of 28Gbits/sec and 20.5Gbits/sec. Random performance is also good, with Iometer recording cumulative read and write rates of 26.1Gbits/sec and 13.5Gbits/sec.

I/O throughput is on the money too, as we recorded total sequential read and write rates of 151,000 IOPS and 67,000 IOPS. For random operations, we saw total read and write rates of 116,000 IOPS and 64,000 IOPS.

Verdict

SMBs that want a top performing storage array at a sensible price won't find a better solution that the new MSA 2050/2052 appliances. The price we've shown is for the 2052 SFF SAN model Q1J03A which includes a pair of 800GB mixed-use SSDs to get you started and with all data tiering features enabled as well, it looks excellent value.

Verdict

The MSA 2052 is a cinch to deploy and delivers excellent performance - a great choice for SMBs that want automated data tiering at a price they can afford

Chassis: 2U rack

Storage: 12 x LFF or 24 x SFF hot-swap drive bays

Drives included: 2 x 800GB mixed-use SAS SSDs

Power: 2 x 570W hot-plug PSUs

Two controllers each with the following:

Memory: 8GB cache with super-capacitor and CompactFlash card

RAID support: RAID1, 10, 5, 6

Data ports: 4 x SFP+ supporting 8/16Gbps FC or 1/10GbE iSCSI

Management: 10/100 Ethernet

Expansion: 6Gbps SAS port (max 96 LFF/192 SFF disks)

Other ports: Mini-USB CLI port

Volume Copies/Snapshots: 128/512

Software: HPE MSA Advanced Data Services Suite

Management: Web browser

Warranty: 3 years parts only NBD

Choice has always been a big feature of Infortrend's EonStor storage arrays and the latest DS 4000 Series has it by the bucketful. These use hybrid controllers supporting four host board options allowing you to choose from 10GbE iSCSI, 8/16Gbps FC, 10GbE FCoE and 12Gbps SAS-3.

The new modular controllers include dual 1GbE iSCSI ports and have room for two host boards. There's more as the controller's 2.2GHz Xeon D-1500 SoC supports DDR4 cache, the standard 8GB can be boosted to a whopping 256GB and it's protected by a fast-charging capacitor.

Snapshots, volume copies and volume mirrors are included in the price while optional data tiering supports up to four storage tiers where hot data is kept at the top for the best performance. Remote replication is also an option as is the new SSD cache feature which uses up to four SSDs per controller to accelerate random read access.

On test is the DS4024RUCB 24-bay model which Infortrend supplied with a full house of 200GB Toshiba SAS-3 eMLC SSDs along. There are dual controllers, each with four 16Gbps FC ports. Expansion potential is staggering as the controllers have dual SAS-3 ports and support up to 444 drives using Infortrend's 16, 24 and 60-bay disk shelves.

Controllers and MPIO

In dual controller arrays, both are active and all host ports are available. However, if you use them for creating redundant links over multi-path I/O (MPIO) links, the path to the secondary controller functions as a standby link and its bandwidth is not aggregated with the primary link.

When you create a logical drive, you assign either controller A or B as its primary one. Essentially, the port on the secondary controller only comes into play if the primary port fails.

This gives you more choice as you can create multiple virtual volumes and use either controller as their primary link. It's also possible to create aggregated MPIO links simply by mapping volumes to multiple ports on the same controller.

For Windows Server 2012 R2, avoid Infortrend's EonPath software and only use Microsoft's DSM for MPIO links. The muddled and outdated user manual fails to mention this and we found that EonPath doesn't provide options to choose MPIO policies such as round robin or weighted paths.

Easy storage provisioning

Deployment is a piece of cake as we used the array's RAIDWatch web console to create two RAID6 logical drives each with 12 SSDs and assigned one to Controller A and the other to B. Within the drives, we created partitions of varying sizes and mapped them to selected FC ports.

Drives are placed in logical volumes with an appropriate RAID level assigned

RAIDWatch only provides basic access to storage services so you'll need to use the SANWatch app to manage snapshots, volume copies, replication and data tiering. Its interface is very similar to the RAIDWatch console so it's easy to use and its home page provides a useful array of performance graphs.

Partition snapshots can be scheduled regularly - as often as every 5 minutes

The price includes a standard snapshot license supporting 64 per volume and 128 per system which can be taken manually or scheduled as often as every 5 minutes. Rollback is fast as we recovered deleted data from a partition in 30 seconds although it did have to be unmapped first.

SSD caches and tiering

As our review system was fully populated by SSDs there was nothing to be gained from the SSD cache feature. For arrays using standard hard disks, you install SSDs in the head unit only and place them in the cache after which the array automatically moves frequently accessed data to them.

For data tiering, you have a 2-tier license costing a one-off fee of 1,899 with a 4-tier license increasing this to 2,999. To use it, you create separate logical drives, enable tiering and assign each a priority from 0 to 3 where Tier 0 would comprise fast SSDs.

Sadly, data tiering doesn't occur in real-time so you must run on-demand or scheduled migration jobs from SANWatch. Scheduled jobs can be run as often as once a day and move blocks to the appropriate tier based on usage.

Different drives can be placed in up to four tiers based on their performance and capacity

Stratospheric performance

To test Infortrend's bold performance claims, we used four Xeon E5-2600 servers running Windows Server 2012 R2 and all equipped with Emulex LightPulse dual-port 16Gbps FC adapters.

We created four partitions and mapped each one to two ports on the same controller so we could create 32Gbps MPIO links. Using Iometer, we recorded around 2,560MB/sec for sequential reads on each server giving us a cumulative 10,240MB/sec not far off the claimed 11,000MB/sec.

We found Infortrend a tad pessimistic on its write speeds as instead of the claimed 5,500MB/sec, we saw a much higher cumulative speed of 7,700MB/sec. General I/O performance was remarkable when using small 4KB transfer requests in Iometer, our four servers generated an incredible 1 million IOPS for sequential reads and 526,000 IOPS for writes.

During our lab tests we recorded over 10,000MB/sec for Iometer sequential read operations

Conclusions

The hybrid controllers in the new EonStor DS4024RUCB make it one of the most flexible storage arrays we've yet seen. It packs in an impressive range of features for the price and has huge expansion potential while performance for all-flash arrays is simply stunning.

Verdict

Its fine blend of features and amazing all-flash performance makes this storage array a top value choice for storage hungry businesses

Chassis: 2U rack

Storage: 24 x SAS-3 SFF hot-swap drive bays

Power: 2 x 530W hot-plug PSUs with cooling fans

Two hot-swap controllers each with the following:

CPU: 2.2GHz Intel Xeon D-1500 SoC

Memory: 8GB DDR4 cache (max 256GB) with capacitor

Array support: RAID0, 1, 3, 5, 6, 10, 30, 50, 60, JBOD, global hot-spare

Host interfaces: 2 x 1GbE iSCSI, 2 x host board slots

Network: 1GbE management port

Expansion: 2 x 12Gbps SAS

Management: RAIDWatch, SANWatch

Options: Dual 16Gbps FC board, £977; dual 10GbE SFP+ iSCSI board, £390; SSD Cache license, £1,499 (all ex VAT)

EMC has announced two new products aimed at moving datacentres away from hard disks and to flash storage. The VMAX All Flash is a series of solid state disk arrays that go all the way up to 4PB or 4000TB.

Available in two new models, the VMAX 450 and 850, EMC claims VMAX All Flash is the biggest all flash storage unit on the planet' yet is less expensive than comparable traditional 15000rpm hard disk arrays. The company also claims that its Write Folding technology improves flash durability without affecting performance.

Each VMAX All Flash array is made of V-Bricks. A V-Brick consists of a single VMAX controller and 53TB of storage capacity that can be increased to 500TB in 13TB steps. Each V-Brick can manage up to eight controllers and 4PB of storage and can cache up to 16TB.

The entire array is backed by a flat-priced maintenance plan the Xpect More Program. This lifetime plan will guarantee the endurance of the flash modules, replacing them for free if they fail. Plus there's a three-year, money-back warranty.

EMC also launched DSSD D5, a PCI-based 144TB 5U flash storage array based on EMC's own Rackscale Flash' silicon design and almost non-existent' latency. Based on industry standard PCI Express Gen3 and NVMe technologies, EMC claims performance of 10 million IOPS, 100GB/s speeds with a 100 microsecond latency.

Up to 48 direct attached servers can use each DSSD D5. Each D5 array consists of 36 custom form factor hot pluggable flash modules. Each module has two NVMe PCI Express-based connectors for use with two controllers simultaneously. The connection cables themselves are 96 thick four-lane leads. It's worth bearing in mind that only' 100TB of the 144TB raw capacity available for use.

EMC claims reliability isn't an issue due to its own proprietary wear levelling, garbage collection and Cubic' RAID technologies. Plus, with redundant components, there shouldn't be a single point of failure and thus it should be well-suited for mission-critical enterprise systems.

To connect it to your servers, EMC sells PCI host adapter cards that plug into any server. To get the most out of the DSSD D5, EMC recommends using its own Flood API which bypasses the file system and other aspects of the operating system to connect applications directly to the SSD.

The US military "fully intends to take advantage of the real time analytics" made possible by DSSD D5 according to Douglas Babb, a chief IT architect for the Department of Defense. Meanwhile, the Texas Advanced Computing Center is already running 50 to 100 scientific supercomputing projects on DSSD D5 according to its executive director, Dan Stanzione.

A revolutionary storage medium, capable of holding 360TB of data and withstanding extreme temperatures, will allow humanity's life and works to survive for "billons of years", its creators say.

Scientists from the University of Southampton's Optoelectronics Research Centre (ORC) have referred to the breakthrough technology, which uses nanostructured glass, as being akin to the memory crystals seen in Superman.

Known as 5D, the medium takes its name from the process of five dimensional digital recording that allows information to be stored on specially created glass by femtosecond laser writing.

The scientists say their ultra-compact technology has a thermal stability up to 1,000C, which at room temperature gives it a "virtually unlimited lifetime".

The prototype shown above measures just 25mm in diameter smaller than a 2 coin (28.4mm). Speaking to IT Pro, ORC team member and PhD student Ausra Cerkauskaite said the researchers would look to produce a standard-sized optical disc (12cm), which would then "ultimately fit about 360TB" more than 7,000 times a 50GB double-layer Blu-ray disc.

"Despite all technological progress it is still difficult to securely store large amounts of information over even relatively short timescales of 100 years. Our solution allows completely eliminating this problem. And this does fascinate us," Cerkauskaite said.

5D could be the start for a new era of colossal data archiving. The team at the ORC believes its technology could be highly useful for national archives, museums and libraries to preserve their records, photos and other documents.

The researchers first demonstrated an experimental version of the technology in 2013 when a 300KB digital copy of a text file was successfully recorded in 5D.

Now, they have managed to write historical and cultural work such as the Universal Declaration of Human Rights (UDHR), Newton's seminal work Opticks, the Magna Carta and the King James Bible to 5D.

"It is thrilling to think that we have created the technology to preserve documents and information and store it in space for future generations," said professor Peter Kazansky, from the ORC. "This technology can secure the last evidence of our civilisation: all we've learnt will not be forgotten."

The researchers are due to present their findings at SPIE Photonics West, an international optical engineering conference in San Francisco, this week.

The team is now looking for industry partners to further develop and commercialise the technology.

Cerkauskaite told IT Pro: "Our main goals now are to improve the speed of writing and develop the technology of microscope-free read-out drive that would increase the impact towards the realisation as mass-market technology."

Images: University of Southampton, Optoelectronics Research Centre

Netgear's rack-mounted ReadyNAS family gets a shot in the arm as its latest 3138 model combines improved hardware along with a heap of data protection features. Its data safety net includes Netgear's versatile X-RAID2 arrays, mixes in copy-on-write data protection with unlimited snapshots and adds private cloud backup plus anti-virus scanning.

Processing power gets a boost as it replaces the 1.7GHz dual-core Atom C2338 from the ReadyNAS 3130 with a nippy 2.4GHz quad-core Atom C2558. Memory also sees a doubling as base systems come with 4GB of ECC DDR3 expandable to 16GB. If you go for a diskless model then we recommend checking Netgear's certified hard disk list. Netgear may decline to provide support if you use drives that aren't on this list.

Installation issues

Our review unit came with four certified 1TB WD Red NAS disks which only took us a few minutes to fit into the hot-swap drive carriers. It went downhill rapidly from here on though - while the RAIDar discovery utility found the appliance on the network, all attempts to create a RAID array from the web console were unsuccessful.

After a call with Netgear support, we found that the third drive had failed. We could check their status by hovering the mouse pointer over the drive icons in the System page and looking at the ATA Error count in the drop-down display. This process could be far more intuitive.

Rather than wait for a replacement to be shipped, we created an X-RAID2 array with the remaining three drives. It's worth using Netgear's X-RAID2 rather than its Flex-RAID option (which is essentially just standard RAID) as the former supports a single volume using different sized drives and will auto-expand capacity when you add new drives or swap them out for larger ones.

It's also a better choice for the 3138 as you can't expand outside the box. You can plug in USB storage devices, share their contents on the network or use them as backup locations but the dual eSATA ports don't support Netgear's EDA500 expansion units.

Netgear's tidy web console provides easy access to all storage features

Data protection plans

When setting up your shared folders and iSCSI LUNs, selecting the bit-rot protection' option enables the BTRFS copy-on-write feature. More importantly, ticking continuous protection' creates regular snapshots to hourly, daily or weekly schedules.

Data protection is simple to set up and the new ReadyOS 6.4 console is more informative than before

This is fully automated and we could hide snapshots or make them visible as new network shares for swift drag and drop file restore operations. Recovering files, folders and LUNs from the web console is just as easy all we had to do was go to the Dashboard's timeline graph, pick a snapshot and select the Rollback option.

Snapshots can also be cloned and made visible as new shares or LUNs. Access security is good as we could enable specific network protocols for each share and decide which users could access them.

Clouds and apps

Netgear's ReadyCloud portal is very useful as once we had registered the appliance with our account, we could manage it remotely from anywhere. We could add other users to ReadyCloud so they could remotely view, add or delete files and folders.

They could copy files to the appliance from their desktop by dragging them into the portal's Browse page. Simple file syncing is achieved using the ReadyCloud agent and Netgear provides versions for Windows and Macs along with apps for iOS and Android.

ReadyCloud provides easy remote access for users plus file synching services

Once the agent has logged in, users can see the shared folders they have permission to use and data in selected local folders is synched with the appliance. Selecting a remote folder in the pop-up agent window also provided direct access to it from Windows Explorer.

You can link the appliance with a Dropbox account, but Netgear still hasn't provided any apps for backing up to Amazon S3 or Glacier cloud storage. Other apps are also thin on the ground and the Surveillance app is inferior to the equivalent app provided by Synology for their range of NAS devices.

Netgear's Surveillance app is useful but can't match Synology's version for features or ease of use

Performance

If you already have a ReadyNAS 3130 and are using it for light storage duties, then you won't see any performance improvements with the 3138. For simple drag and drop copies of a 25GB test file, both the 3130 and 3138 returned average read and write seeds of 113MB/sec and 112MB/sec.

They both handled our backup test well with a 22.4GB test folder and its 10,500 small files copied at 87MB/sec. However, we found the 3138 handled heavy loads better.

We connected four Xeon E5-2600 v3 rack servers and mapped shares over dedicated links to the 3138's four Gigabit Ethernet ports. Running Iometer on each server saw cumulative read and write speeds of 452MB/sec and 288MB/sec respectively. The same test on a 3130 returned moderately lower cumulative speeds of 430MB/sec and 260MB/sec.

There's nothing to be gained with the 3138 if you're planning on using Netgear's performance-sapping Anti-Virus app for real-time scanning. With this enabled, our 22.4GB backup test folder was copied down at 32MB/sec the 3130 mustered 30MB/sec for the same task.

Conclusions

At 610 for a diskless model, the ReadyNAS 3138 is around the same price as Synology's RS815+. This delivers equally good performance, but wins out with its superior feature set and external expansion capabilities.

However, with its unlimited snapshot feature, the 3138 is the better choice if data protection is your top priority. Snapshots are really easy to set up, data recovery is a cinch and Netgear's extensive cloud services makes it very accessible as well.

Verdict

A great network storage choice if you want simplified data protection and good cloud access – plus its uprated hardware won’t wilt under pressure either

Chassis: 1U rack

CPU: 2.4GHz Intel Atom C2558

Memory: 4GB ECC DDR3 UDIMM (max 16GB)

Storage: 4 x hot-swap 2.5in/3.5in SATA drive bays

Array support: RAID0, 1, 5, 6, X-RAID2, JBOD

Ports: 2 x USB 3, USB 2, 2 x eSATA

Network: 4 x Gigabit Ethernet

Warranty: 5 years with NBD hardware replacement

Prices for 10-Gigabit (10GbE) networking kit are dropping rapidly making it much more affordable for SMBs. This hasn't gone unnoticed by most NAS appliance vendors with Qnap the first to include a 10GbE SFP+ fibre adapter with a NAS, specifically the TS-x63U series.

Another feature that makes this family stand out is its wholesale adoption of AMD's 2GHz GX-420MC SoC (system on chip). This low-cost embedded x86 processor has four cores running at 2GHz, supports DDR3-1600 memory and has a low TDP of 17.6W.

The flagship TS-1263U-RP on review is a 2U rack mount system with 12 hot-swap SATA drive bays and dual hot-plug PSUs. You can add only a single external disk shelf, either the eight-disk SATA UX-800U-RP or the 12-disk UX-1200U-RP. Qnap's 10GbE card nestles in the single PCI-Express slot but it doesn't come with a transceiver so you'll need to source one separately - the Avago 850nm models we use cost around 80 ex VAT.

Build quality and deployment

The appliance exhibits the classy build quality we've come to expect from Qnap with a solid steel chassis and sturdy metal drive carriers. The lid is easily removed for upgrades and the dual internal cooling fans are very quiet.

The appliance is supplied with 4GB of DDR3L memory on a single SO-DIMM module. You can fit up to 16GB, but as there are only two memory slots you'll need to dispense with the pre-installed 4GB to max it out.

Installation is deftly handled by Qnap's cloud portal as we entered the unique cloud key on the label on the lid and followed the quick start wizard which installed the latest QTS firmware. We also popped in a quartet of 4TB WD Red NAS Pro SATA drives and let it create a RAID5 array from them.

The discovery web portal helps set up storage but make sure you choose thick or thin volumes if you want snapshots

New storage features

Qnap's latest QTS 4.2 firmware has a wealth of data backup features including excellent support for a wide range of cloud storage providers. We successfully backed up folders on the appliance to our Amazon S3 account using the Backup Station app, but Qnap also provides free apps for Amazon Glacier, Microsoft Azure, Google Cloud Storage and many more.

Qnap offers an impressive range of backup apps and supports a multitude of cloud providers

Workstation backup gets a boost as the new Qsync Central Station app allows users to link up with authorised shares on the appliance and synchronise folders with them. The Qsync Windows app is easy to use and for swift file restoration, it provides a direct link to the browser-based File Manager.

The File Manager also gets a makeover and now supports direct links to cloud storage providers. We linked it with our Dropbox and Google Drive accounts but it also supports remote links to FTP servers and allows folders to be remotely mounted using WebDAV.

Snapshots

No self-respecting NAS vendor would now be seen without snapshots and QTS 4.2 supports them for both thick and thin multiple volume types (but not static volumes). You can run them on-demand or schedule them regularly for NAS volumes or iSCSI LUNs. You use the Snapshot Manager app to keep track of them.

Scheduled smart snapshots can save storage space as the appliance will only take the next snapshot if the data has changed since the last one. The Snapshot Manager makes light work of data recovery. We could view snapshots and use the Revert option for full recovery, or view NAS share contents and restore individual files.

The new Windows Snapshot Agent can be used to take application-consistent snapshots on iSCSI LUNs. Whenever the appliance wants to take a LUN snapshot, it'll inform the agent first which instructs the host system to flush all data from memory into the LUN.

Snapshots can be scheduled and Qnap's Windows agent makes sure they're application-consistent

Good management and performance

If you have multiple Qnap appliances then you can use the new Q'centre app to centrally manage and monitor them all from a single console. We used custom dashboards to create views of the appliance with graphs of storage pool and volume usage plus appliance hardware utilisation.

Qnap now offers Q'centre as VMware and Hyper-V virtual appliances. We tested the latter on our Windows Server 2012 R2 Hyper-V host. After importing the free VM, we found it delivered all the same features as the on-appliance version.

We used the new Q'centre Hyper-V VM to centrally manage the lab's Qnap appliances

For performance testing, we mapped a share over 10GbE to an HP ProLiant DL380 Gen9 rack server running Windows Server 2012 R2. Iometer reported fast raw read and write rates of 9.2Gbits/sec and 5Gbits/sec.

Real world NAS speeds are good too, with drag and drop copies of a 25GB test file returning read and write speeds of 412MB/sec and 316MB/sec. Backup operations are reasonably fast with our 22.4GB test folder and its 10,500 small files copied to a share at an average of 183MB/sec.

Conclusions

Pricewise, the TS-1263U-RP is a winner as even though you'll need to source a 10GbE transceiver, it's still excellent value. Synology's 12-bay RS2416RP+ for example, costs nearly the same but has no PCI-Express slots and so can't be upgraded to 10GbE.

And then there's Qnap's own TS-1280U-RP which comes with a 3.4GHz E3-1200 v3 Xeon, dual PCI-Express slots and a massive expansion potential but costs around 800 more. If you're looking for a 10GbE-ready network storage appliance with a balanced combination of features, performance and price, then the TS-1263U-RP is well worth considering.

Verdict

Limited external expansion but this 12-bay appliance scores highly for storage features, its 10GbE readiness and its price

Chassis: 2U rack

CPU: 2GHz AMD GX-420MC SoC

Memory: 4GB DDR3 (max 16GB)

Storage: 12 x 2.5/3.5in SATA hot-swap drive bays

Array support: RAID0, 1, 5, 6, 10, hot-spare, JBODs

Network: 4 x Gigabit Ethernet; 1 x LAN-10G1SR-D 10GbE SFP+ card

Other ports: 2 x USB 3, 2 x USB 2

Power: 2 x 250W hot-plug PSUs

Management: Web browser

Warranty: Two year limited

Qsan has a reputation for delivering affordable IP SAN arrays with high expansion potential and it has done it again with its latest AegisSAN Q500-P20-D424. This imposing 4U system gets you started with 24 SAS/SATA hot-swap bays and can be expanded up to 256 drives using Qsan's J300Q external disk shelves.

SMBs will like its price tag as our diskless review system with dual 10GbE-enabled controllers costs just 6,328 ex VAT. Even better, both controllers function in active/active mode and provide fast, transparent failover.

The P20 controllers each provide two 10GbE SFP+ and dual Gigabit Ethernet data ports along with dedicated Gigabit management ports. To cut costs, you can start with one controller and upgrade later on. Qsan also offers models with six Gigabit Ethernet, dual 8/16Gbps FC or dual 6Gbps SAS ports.

Top tolerance

Hardware fault tolerance is excellent as the array doesn't have a single point of failure. You get triple 550W hot-plug PSUs, plus dual cooling fan modules. Qsan also offers an optional cache memory battery backup pack.

Both controllers have 6Gbps SAS expansion ports so the J300Q disk shelves can be daisy-chained over dual links. These also have dual hot-plug power supplies and cooling fans.

You don't have to swap out the controllers if you want different data ports as they can be upgraded in the field. The kit comprises a daughtercard, SATA DOM plus a new backplate. We upgraded a P10 Gigabit model to the P20 10GbE version on review in around 30 minutes.

We had no problems deploying the array in our lab as Qsan's QFinder located it for us and provided direct access to its web interface. We also used Qsan's QCentral Java app which can manage multiple AegisSAN and TrioNAS appliances from a single interface.

Qsan provides plenty of quality management tools for the array

Disk and RAID support

For testing, we loaded up four 4TB WD Enterprise SAS drives. For SATA drives, you'll need Qsan's MUX interposer boards. We found them easy to fit on the back of the drives carriers, but their price tag of 30-40 each does nullify the cost savings of cheaper SATA drives.

Storage provisioning is easy enough and we started by creating a RAID Group (RG) using all our SAS drives and choosing an array type. Qsan supports stripes, mirrors, RAID10, 5, 6, 50 and 60 or you can use the wizard which grabs all the drives and chooses an array type for you.

Thin provisioning can be applied to selected RGs and policies used to handle space usage and alerting. Each policy employs six user-defined capacity thresholds that can trigger actions such as an email message, space reclamation, snapshot deletion and, when no more space is available, they can safely take the RG offline. If space gets tight you can install extra drives and add them to the RG on the fly.

The array's local web console provides full access to all storage features

VDs and performance

IP SAN targets are swiftly provisioned by creating virtual disks (VDs) within an RG and choosing a capacity for each one. When new drives are added to an RG, the VD can be expanded into them.

We attached each VD to a LUN and could decide which of our host initiators were allowed to see them. Global access is enabled simply by using a wild card and we could set each LUN as read only or read/write.

To test performance, we used an HP ProLiant DL380 Gen9 rack server equipped with an Emulex dual-port 10GbE adapter and running Windows Server 2012 R2. With it logged in to a 750GB target, we saw Iometer report good raw read and write speeds of 1,060MB/sec and 809MB/sec which equates to 8.3Gbits/sec and 6.3Gbits/sec.

With a dual 10GbE MPIO link across both controllers, we tested failover by firing up Iometer and then pulling out the primary controller. The test continued unabated (albeit at the speed of a single link) and when we plugged the controller back in, the array spotted it and enabled full redundancy in less than 60 seconds.

We could keep a close eye on performance from the QCentral console

Snapshots and clones

We used the web interface and QCentral to configure snapshots for selected VDs, run them on-demand and schedule them for 15-minute, hourly, daily, weekly and monthly intervals. Both interfaces provide lists of existing snapshots which can be exposed as new read-only or read/write targets and assigned to specific hosts using the handy auto-map feature.

Data recovery is a cinch. To test this, we deleted all the data in a 500GB target and dismounted it at the host. With the latest snapshot selected, we chose the rollback option from its drop-down menu and had all the data restored in less than 5 seconds.

Clones are just as easy to use as you create a Backup VD that's the same size as the source. When it's ready, select the Set Clone option on the source, choose the Backup VD as the target and either run the job manually or apply a schedule to keep the clone regularly updated with the latest snapshots.

Snapshots can be run on-demand and scheduled at regular intervals

Conclusions

The Q500-P20-D424 delivers fast performance, all-round redundancy and huge expansion potential. If that's not enough, it also has dual active controllers and fast transparent failover at a price the blue chips will have a tough job competing against. It's superb value.

Verdict

Qsan’s IP SAN array combines big capacity with enterprise features at an SMB price. Dual active controllers, fast failover and total redundancy make it perfect for storing business critical data.

Chassis: 4U rack

Storage: 24 x hot-swap LFF/SFF SAS/SATA drive bays

Power: 3 x 550W hot-plug PSUs

Cooling: 2 x hot-plug fan modules

Dual active/active RAID controllers each with:

Processor: 2GHz Atom S1289 SoC

Memory: 4GB DDR3 with battery backup module (max 8GB)

Array support: RAID0, 1, 10, 3, 5, 6, 10, 30, 50, 60

Network: 2 x Gigabit Ethernet, 2 x 10GbE SFP+, 1 x management

Other ports: SAS expansion port

Management: Web browser, QCentral

Fujitsu's storage division has traditionally focused on mid-sized businesses and enterprises, but it now wants a piece of the SMB market. The Eternus DX60 S3 IP SAN array delivers all the key features of its bigger brothers but at a very affordable price.

The 2U chassis is available with 12 LFF or 24 SFF drive bays along with single or dual controllers. For host data ports, Fujitsu offers controllers with dual 4/8Gbps Fibre Channel, Gigabit Ethernet or 6Gbps SAS.

The appliance supports SAS and NL-SAS drives, but only 6Gbps speeds are supported. Furthermore, if you want faster data port speeds such as 10GbE then you'll need to upgrade the system with Fujitsu's DX100 S3 controllers which can be done in the field.

Expansion options are limited as the DX60 S3 only supports a maximum of 24 hard disks, so the SFF model can't be expanded as it's maxed out already. The 12-bay LFF model supports one expansion shelf allowing maximum capacity using certified 6TB NL-SAS drives to be pushed to 144TB.

Deployment

Initial installation is simple enough as the appliance's web interface provides a quick start wizard. This requests details such as a hostname, a new administrator password, SNMP community names and license registration.

Next, we created a RAID group which defines a collection of drives, a total capacity and the primary controller that will look after it. Multiple data volumes are created within each RAID group and for each one, we could choose from stripes, mirrors, RAID10, 5, 6 or 50 arrays.

RAID groups can be migrated from one array type to another and volumes expanded into spare space. Thin provisioning requires a dedicated storage pool and as volume space gets used, the appliance dynamically assigns new blocks to it from the pool.

The appliance's web console keeps you posted on the status of RAID groups, thin provisioning and SDVs

Affinity groups

It gets more complex from here on as before mapping volumes to our host systems, we needed to create host, data port and LUN (logical unit number) groups first. Multiple volumes can be assigned to groups where they each get their own LUN number.

Host servers are placed in groups but their initiators need to be logged in to the appliance's data ports first. The wizard provides an auto-discover feature, but it failed to locate our hosts so we had to enter their IQN details manually.

With our host, port and LUN groups created, we could now use Fujitsu's affinity groups to link them all together. It's a lengthy process, but does provide tight access controls as hosts can only access selected volumes via the assigned data ports.

Affinity groups bring hosts, LUNs and data ports together and are used for strict access controls

Performance

To test performance, we used the lab's HP ProLiant DL380 Gen9 rack server running Windows Server 2012 R2. After manually adding it as a host, we used an affinity group to link it up with a RAID5 volume accessible from data ports on both controllers.

With Iometer configured for 256KB transfer requests, we saw raw sequential read and write rates for a 500GB target of 113MB/sec and 112MB/sec. With a dual-port MPIO link to the same target, we saw speed double to 226MB/sec and 224MB/sec.

The appliance's web interface provides a good range of performance monitoring tools. From the Component tab, we could view bar graphs of physical disk utilisation and move to the Volume tab to see read and write IOPS plus throughput in MB/sec for each volume.

Data protection features

The appliance provides a wealth of snapshot, cloning and backup features, but most of them require the Eternus SF Express management server which is installed on a separate Windows host system. This connects to the appliance remotely via SNMP and provides access to Fujitsu's Advanced Copy services.

Fujitsu's Eternus SF Express web server is used to manage multiple appliances and IP SAN hosts

Basic volume cloning is achieved using the one point copy (OPC) feature which takes a source volume and creates a single copy of it. QuickOPC takes an initial volume copy and updates it incrementally while Equivalent Copy (EC) manages mirrors of one volume to another on the same array.

SnapOPC+ (one point copy plus) take point-in-time copies, or snapshots, of selected source volumes but requires dedicated snap data volumes (SDVs) to be created on the appliance. These can be used for offloading backup as once the SnapOPC+ task has finished, the SDV can be mapped to a backup server where it appears as a perfect copy of the original.

Remote Equivalent Copy (REC) mirrors volumes from one array to another, but isn't supported by the DX60 S3. Other options also not available are Fujitsu's optional data tiering and support for NAS services.

Advanced Copy snapshot and cloning features are also configured remotely from the SF Express web console

Conclusions

The Eternus DX60 S3 is inexpensive as we were quoted a price of 3,749 ex VAT by Ebuyer.com for a diskless SFF chassis with dual redundant Gigabit Ethernet-equipped controllers. It's fast, but small businesses could find the deployment processes and management software overly complex. Fujitsu advised that it provides engineer-assisted installation and training services which we would strongly recommend taking advantage of if you want the DX60 S3.

Verdict

Dual controllers and plenty of storage features make this a good IP SAN disk array, but expansion is limited and its data protection services are all handled by a separate management utility

Chassis: 2U rack

Storage: 24 SFF hot-swap drive bays

Power: 2 x 800W hot-plug PSUs

Dual hot-plug controllers each with the following:

CPU: 1.2GHz Pentium

Memory: 2GB DDR3

Array support: RAID0, 1, 10, 5, 6, 50, hot-spare

Data ports: 2 x Gigabit Ethernet

Expansion: SAS port (only supported on LFF chassis)

Management: Web browser

Software: Fujitsu Eternus SF Express

The RC18015xs+ is a radical departure from the norm for Synology as this is its first fault tolerant NAS and IP SAN storage solution. Aimed at cost-conscious SMBs, the standard configuration uses two RC18015xs+ controllers which function as an active/passive failover cluster.

The appliances are linked together via dedicated heartbeat ports so if the active controller goes down, the passive unit steps up to the plate and takes over. The controllers each have a SAS expansion port where up to fifteen 12-bay RXD1215sas disk shelves are linked to both of over redundant paths allowing raw capacity to reach well beyond 1 petabyte.

Synology makes the jump to BTRFS and a key feature of this file system is unlimited block-level snapshots. Netgear made a wholesale move to BTRFS last year but a big difference is Synology retains support for traditional EXT4 volumes.

The cluster controllers support up to fifteen RXD1215sas disk shelves for massive expansion

Cabling and deployment

To create a cluster from two RC18015xs+ appliances, we linked their Gigabit heartbeat ports using a standard Cat6 cable. Each link to the production network requires the corresponding ports on both controllers to be connected and we added a single RDX1215sas disk shelf to both of them using the supplied SAS cables.

Powering both units up automatically turns on any attached disk shelves and we then moved over to Synology's standard discovery web portal for further configuration. This found the cluster and installed the latest DSM software for us.

We loaded a quartet of 4TB WD SAS drives and created a single RAID5 array. We could add extra drives as hot spares or expand the array into them but this solution doesn't support SATA drives.

The DSM Storage Manager shows the physical connections between the controller and disk shelves

Cluster control

The DSM web console is virtually the same as used by Synology's standard NAS appliances. The only differences are the new High Availability Manager (HAM) and Data Protection Manager (DPM) apps.

The HAM provides an animated status graphic for the cluster and attached disk shelves. From here, we could switch the active and passive assignments, reboot and power either unit off or shut down the entire cluster.

The standard DSM Storage Manager app is used to create RAID disk groups, volumes, IP SANs and SSD caches. It also provides an extra graphic showing the physical connections between the disk shelves and controllers.

The new High Availability Manager app keeps you posted on cluster status

Snap happy

Volume creation using BTRFS is simple and during shared folder creation, we enabled protection and decided whether their snapshots should be visible. From the DPM app, we created manual and scheduled snapshots for selected shares and decided how many hourly, daily, weekly, monthly and yearly snapshots to retain.

The same procedures apply to iSCSI LUNs and Synology will also be releasing a Snapshot Manager for taking application consistent snapshots in Windows and VMware environments. Entire shares and LUNs can be restored from DPM by selecting the desired snapshot and starting the recovery process.

We deleted a 50GB folder on one share and restored it from the latest snapshot in under 20 seconds. Visible snapshots show all data natively so we could browse the network share and restore files and folders using drag and drop.

BTRFS support allows you to schedule unlimited snapshots for NAS shares and iSCSI targets

Failover and performance

To test failover, we powered off the active controller from the HAM app whereupon it flagged up a warning message that all storage services would be interrupted while it promoted the passive controller. The process took 90 seconds and during this phase, we lost contact with all our network shares plus iSCSI LUNs and were logged out of the DSM console.

Once the controller had been promoted, everything sprang back to life but apps such as backup will be terminated if the active controller fails. We tested this by running Iometer on a share during this process and it had to be restarted when failover had finished.

There was nothing to do when the failed controller was powered back up. It re-established contact with the new active controller, was automatically assigned the passive role and all without any further interruptions to storage services.

For 10GbE testing, we installed dual-port Emulex fibre SFP+ cards in both controllers and linked them to a Netgear 10GbE switch. With a share mapped to an HP ProLiant DL380 Gen9 server (web ID:23173), we saw Iometer report raw read and write speeds for a network share of 1,108MB/sec and 1,100MB/sec.

Real world copies of our 50GB test file also returned fast sustained read and write speeds of 412MB/sec and 400MB/sec. IP SANs are fast as well with Iometer returning raw read and write speeds for a 500GB target of 1,060MB/sec and 1,000MB/sec.

Failover worked fine but it will interrupt all storage services during this brief phase

Conclusion

The RC18015xs+ puts Synology into a whole new market and the minimum hardware of two controllers and one disk shelf is comparatively expensive. Qsan's TrioNAS LX U400HA-D424 (web ID:24575) costs around the same but offers dual active/active controllers and fast, transparent failover for NAS and IP SAN services.

That said, with Synology's DSM software at the helm, the RC18015xs+ delivers an impressive range of storage features you won't find elsewhere. Support for BTRFS volumes also provides valuable unlimited snapshots, it delivers good performance and expansion potential is huge.

Verdict

The RC18105xs+ performed well during testing and delivered very good NAS and IP SAN performance. Failover is nicely automated but it will interrupt services and for an active/passive fault tolerant solution, it isn’t the best value

2 x RC18015xs+ each with the following specification: Chassis: 1U rack CPU: 3.3GHz Intel Xeon E3-1230 v2 Memory: 8GB DDR3 (max 32GB) Network: 5 x Gigabit (4 x LAN, 1 x Heartbeat) Expansion: 1 x SAS out, 1 x PCI-Express 8X slot Array support: RAID0, 1, 10, 5, 6, hot-spare, JBOD, SHR Power: 2 x 150W hot-swap PSUs Management: Web browser Warranty: 5-year limited RXD1215sas drive shelf (max 15): Storage: 12 x SAS 6Gbps hot-swap drive bays Ports: 4 x SAS (2 x In/2 x Out) Power: 2 x 500W hot-swap PSUs Warranty: 5-year limited

HP has introduced a swathe of new products to push forward the move to all-flash enterprise data centres, including new StorServe 2000-family arrays, as well as a significant cost reduction.

HP 3PAR StorServ 20850 and 20800 are the latest all flash and converged flash arrays respectively. The 20850 can, according to HP, is able to deliver over 75 GBps of sustained data throughput and can offer performance of over 3.2 million IOPS at sub-millisecond latency.

StorServ 20800, meanwhile, offers flash-first converged infrastructure, meaning it is fully flash when installed, but that spinning disk can be added to it, and can scale up to 15PB.

As well as new hardware, HP is trying to tempt in customers with a drop in price from $2 per gigabyte (usable) to $1.50 per gigabyte.

Vish Mulchand, director of product management at HP Storage, told IT Pro that while flash, like any technology, is becoming faster, better and cheaper as it matures and this is part of the reason for the price drop, there are two other influencers - Adaptive Sparing and data compaction technology.

With Adaptive Sparing, HP has worked with drive manufacturers to reduce the amount of memory each of them reserves for control and management of the media.

"Normally, when you buy an SSD drive of 400GB, if you count the capacity in there it is actually 512GB, but the manufacturer keeps some space reserved. When we get a 400GB drive, we also build some spare capacity in it. When we looked at this, we thought 'there is no reason why we and the drive manufacturer both have to build in spares', so we told them to take a smaller amount and we also built in a smaller amount, which means we can essentially make a 400GB drive a 480GB drive. So we give customers 20 per cent free with this technology," said Mulchand.

Data compaction technology, meanwhile, uses HP's Thin Technologies to allow customers to implement deduplication on primary storage, which Mulchand said had "previously been controversial". Due to the efficiencies introduced with this technology, customers can save up to 50 per cent on the effective cost of storage, he said.

"When you combine cheaper drives, Adaptive Sparing and deduplication, all these equate to a 73 per cent drop in prices over the past 18 months and an 8x density improvement," he said.

ServStore 20800 and ServStore 20850 will be generally available from August.

Infortrend has a reputation for delivering low-cost storage arrays to SMBs and its latest EonStor DS1024B takes this to the next level. A key feature of the new DS1000 family is they offer a hybrid storage solution out of the box as their integral Gigabit IP SAN ports can be partnered with optional 10GbE iSCSI or 16Gbps Fibre Channel modules.

There's more as the 24 SFF drives bays support SAS, NL-SAS and SATA drives plus SSDs and InforTrend has added optional data tiering as well. This allows you to create pools containing up to four storage tiers where hot data is kept in the top tiers for the best performance.

End-to-end redundancy is present and correct as the price we've shown includes dual, active/active controllers. These each have 2GB of cache memory protected by large battery backup packs but you can opt for fast-charging super-capacitors instead.

The DS1024B comes as standard with dual, hot-plug PSUs with combined cooling fans. Expansion into Petabyte levels is possible as the controller's 6Gbps SAS expansion ports can be used to daisy-chain extra storage shelves via redundant links with up to 360 SFF drives supported.

Single arrays can be managed using RAIDWatch but this only provides basic levels of access

RAIDWatch or SANWatch?

We found initial installation simple enough as the primary controller's Gigabit management port defaults to DHCP. Only the port on the primary controller is active but if it fails, the array shifts its IP address over to the secondary controller.

The embedded RAIDWatch web server only provides access to the array it's running on. From its tidy web interface, we could create logical volumes from selected drives, manage host channel settings and assign host mappings.

The SANWatch utility manages multiple arrays from the same interface and is required for managing tasks such as snapshots, volume copies and replications. It presents a shallow learning curve as its interface is very similar that of RAIDWatch.

Either way, logical drive creation is swift as we picked our drives, chose a RAID array type and assigned it to a primary controller. Partitions are created within logical drives and then mapped to the data ports on selected controllers.

SANWatch manages multiple appliances and is required to configure features such as snapshots and tiering

Port choices and redundancy

The controllers have four fixed Gigabit data ports and you can choose additional modules to fit in their expansion slots. We were supplied with dual 10Gbase-T and dual 16Gbps FC modules which we found simple enough to install.

Although both controllers are active and their ports all available, you can't actually use them all. During logical drive creation you need to designate either A or B as the primary controller which allows you to use all its ports for MPIO links if required.

The secondary controller sits in the background and provides redundant links. These are assigned to MPIO links but bandwidth is not aggregated with the primary link as it's only used if the latter fails or is taken offline.

Selected drives are placed in logical volumes with an appropriate RAID level assigned

Stonking performance

For 10GbE iSCSI and 16Gbps FC testing, we used a Broadberry Data Systems rack server with dual 14-core E5-2683 v3 Xeons, 64GB of DDR4 plus mirrored SSDs and loaded with Windows Server 2012 R2. We used its dual embedded 10GbE ports for IP SAN tests and a QLogic QLE2672 PCI-e card for 16Gbps FC tests.

The DS1024B delivered top results over 16Gbps FC with Iometer reporting raw sequential read and write rates of 12.4Gb/sec and 11.4Gb/sec. Using small 4KB transfer requests saw a tidy 99,000 IOPS for read operations.

Simply connecting the second FC port on the primary controller to the server automatically created an MPIO link to the mapped RAID5 partition. This returned superb raw read and write rates of 25Gb/sec and 14.6Gb/sec. while I/O throughput ramped up to 148,000 IOPS.

IP SAN performance wasn't as impressive but still very fast nonetheless. With a single 10GbE link to a 100GB target, Iometer reported top raw read and write speeds of 8.6Gb/sec and 7.6Gb/sec plus 90,000 IOPS. With a dual-port MPIO link in the mix, we saw speeds increase to 14.1Gb/sec and 8.6Gb/sec while I/O throughput improved to 115,000 IOPS.

Data tiering can use up to four tiers and migration tasks move data blocks based on age and usage patterns

Snapshots and tiering

The price includes a standard snapshot license which supports 64 per volume and 128 per system. You can increase these with an optional advanced license which increases the count to 256 and 4,096 respectively.

You have two upgrade choices for data tiering where a 2-tier license costs around 1,750 and a 4-tier license costs 2,745. It's easy enough to configure as we created a volume containing logical drives with SSDs set to Tier 0 and SAS drives assigned as Tier 1.

Infortrend's tiering isn't dynamic but is very similar to that used by Fujitsu's Eternus products. The array analyses volume usage and data age and moves blocks accordingly when you run manual or scheduled migration tasks.

Conclusion

The EonStor DS1024B packs in a heap of features for the price and matches them with blue-chip beating FC SAN performance. The optional data tiering may be a manual affair but this enterprise level feature is not normally seen at this low price and its massive expansion potential makes this array a good long-term investment.

Verdict

The EonStor DS1024B is ideally suited to SMBs that want an affordable and highly expandable hybrid storage array. The optional data tiering adds extra levels of versatility and its superb performance over 16Gbps Fibre Channel won’t be bettered by anyone at this price point.

Chassis: 2U rack

Storage: 24 x SATA/SAS SFF hot-swap drive bays

Power: 2 x 460W hot-plug PSUs with cooling fans

Two hot-swap RAID controllers each with the following:

Memory: 2GB cache (expandable to 8GB) with battery backup

Array support: RAID0, 1, 3, 5, 6, 10, 30, 50, 60, JBOD, global hot-spare

Host interfaces: 4 x Gigabit iSCSI, module expansion slots

Network: Gigabit management port

Expansion: 6Gbps SAS port

Management: RAIDWatch web browser, SANWatch Suite

Options: Dual 16Gbps FC module, £872; dual 10GbE module, £598 (all ex VAT)

Mozilla has ditched Google as its default search engine after a 10-year agreement between the companies came to an end.

Firefox will now use Yahoo as its default search engine in the US from December. Under the terms of the five-year deal, Yahoo will support Do Not Track (DNT) within Firefox.

Google, Bing, DuckDuckGo, eBay, Amazon, Twitter and Wikipedia will remain as alternate search options for users who want to manually switch.

In Russia, Yandex Search will be the default search provider and Baidu will remain the defacto engine in China.

The global partnership between Google and Mozilla has not been renewed either but the popular search engine will still be pre-installed.

"Google has been the Firefox global search default since 2004. Our agreement came up for renewal this year, and we took this as an opportunity to review our competitive strategy and explore our options," noted Chris Beard, Mozilla CEO in a blog post.

"We are ending our practice of having a single global default search provider. We are adopting a more local and flexible approach to increase choice and innovation on the web, with new and expanded search partnerships by country."

Mozilla claims its Firefox browser accounts for 100 billion web searches per year - so this looks like a significant coup for Yahoo.

Violin Memory claims its latest piece of software will make it easier for enterprises to build all-flash datacentres that run their business critical apps.

Runs on its Concerto 7000 All-Flash Array, the vendor is rolling out its refreshed enterprise data services software across the globe, today.

The software allows users to take advantage of thin provisioning, asynchronous and synchronous replication, and to generate storage snapshots.

It is claimed the product delivers a lower cost-per-transaction for demanding workloads, resulting in faster application response times, and takes up less floor space than competing arrays.

The 280TB array is designed for use with tier 1 and tier 2 business applications, and according to the company's chief marketing officer Eric Herzog paves the way for enterprises to switch to all-flash storage-based datacentres.

The benefits of doing so, he told IT Pro, include lower power and cooling costs, as well as greater floor space because all-flash arrays can provide a higher performance in a smaller amount of space.

However, despite the relatively higher cost of flash storage, compared to tape or disk, the technology's ability to let users get important work done faster and, therefore more cheaply, is at the heart of the all-flash datacentre's appeal.

"Tape is still the cheapest media, but if you look at it from the holistic perspective of a CIO or IT director, you're not just looking at the cost per gigabyte [anymore], you've got to look at all the costs associated with running a datacentre," he said.

"[When it comes down to the cost of storage] itself, it has traditionally been over-provisioned because you can't get the performance without doing that. Flash takes that need and cost away."

The array is available across the globe from today, and the company has also brought out an upgrade kit so that users of its older 6000 family of all-flash arrays can benefit from the functionality of the enterprise data services software.

The IT Press tour is a five-day event taking place this week in San Francisco that seeks to introduce the UK media to tech firms doing some interesting things in Silicon Valley. IT Pro's Jane McCallion has joined the fray and here's her round-up of the third day's events...

Day three: Would you like optimisation with that technology? Because if so, there are some people here in Silicon Valley who want to talk to you.

Whatever it is you have, you can make it better. At least, that is the message from the first two companies IT Pro met with today in Mountain View and San Jose, California.

The third day of the IT Press Tour started with a visit to software-defined storage vendor Atlantis Computing.

Founded in 2006, the company specialises in virtual desktop infrastructure (VDI) optimisation and aims to make virtual desktops perform better than native desktops.

Put simply, its software allows IT administrators to use RAM as the primary storage for virtualised environments, regardless of the hypervisor they use. According to the organisation, this means businesses can use much less traditional storage, reducing capital expenditure significantly.

"There's no need for businesses to buy SSDs [when Atlantis' software is deployed], because it's all done in the RAM," Seth Knox, VP of products at Atlantis Computing.

"It's the lowest cost desktop out there, [but] it's [also] the best performance desktop out there," added CEO Barnard Harguindey.

Toby Coleridge, the company's global architect, demonstrated the product in action on his MacBook Air by running a script that opened every Microsoft Office application he had installed, including Excel, Word and Access, as well as a number of web pages and other applications simultaneously.

IT Pro can confirm it took seconds to load all the programmes and pages. Coleridge claimed - even on a MacBook Air - it would normally take a couple of minutes.

The Atlantis team also listed some of its clients, including University Hospitals of Leicester, UBS, and 118 118. Financial services firm JP Morgan Chase is also a customer and in the wake of Hurricane Sandy, used Atlantis' technology to replicate virtual desktops and get employees back to work, even though its New York datacentre had been destroyed by flood water.

Nimble Storage: Next on the trail was Nimble Storage, a business specialising in flash-optimised hybrid storage solutions.

The firm recently celebrated its fifth anniversary, but only began shipping in 2010, and has recently moved to new offices in San Jose.

The company is going to have an IPO in the next 12 months (a registration statement was filed in October), which meant it wasn't particularly forthcoming about its business plans and the wider market.

Nevertheless, Nimble has a clear vision to optimise your storage, irrespective of the hardware powering it.

"Flash and disk are complementary," said VP of worldwide marketing Dan Leary.

While flash storage is faster for runtime, he explained, spinning disk still has significant advantages when it comes to cost, sequential read/write and longevity.

Furthermore, Leary said Nimble Software's hybrid system of snapshot and replication, using spinning disk to carry out the operations it is best at and SSD the operations it is best at, making backup and replication virtually instant and more secure.

Additionally, because it is a hybrid offering, administrators can add and/or remove arrays non-disruptively by using Nimble, Leary claims.

The organisation was also able to disclose that in Q3 2014 (i.e. the fiscal quarter ended October 2013), it had over 2,000 clients, up from 780 at the same time last year and 174 the year before that.

Coho Data: The third and final stop of the day was Coho Data in Sunnyvale.

Coho was founded by three men who came together at XenSource, before its acquisition by Citrix, and has been in stealth mode' for two years. It launched in October 2013 and bagged series B funding a month later.

It is early days for Coho, but the team there has a clear vision of what they want to do.

"Coming out of Xen, we realised we didn't want to build yet another storage array. If we had done that, we would have failed," said CEO and co-founder Ramana Jonnala.

Instead, according to CTO Andrew Warfield, who worked on Xen at the University of Cambridge, the company decided enterprises should be able to buy virtualised storage in the way that Amazon does, i.e. in a very commoditised way.

Warfield made plain that in its own view, "[Coho Data] is a software company" and at its heart is a software-defined storage offering. But, self-perception aside, it currently ships this as part of a two-rack unit (2U) "MicroArray", featuring a VMware-facing NFS server. This is because it is, as previously mentioned, early days and integration with other virtualisation systems will be forthcoming.

Tomorrow, Jane will be visiting Soonr and CloudByte.

EMC has reaffirmed its commitment to Flash storage and talked up converged infrastructure and big data as its big bets for the future at its annual EMC Forum event in London.

"We think Flash storage is going to...and already is having a profound impact [on] the storage industry," said Jeremy Burton, executive vice president of product operations, solutions and marketing, during his keynote yesterday.

"Why is that? ... You get dramatic reductions in latencies. If you look at your average performance hard drive, you've got a couple of milliseconds in terms of latency. With a PCIe card in the server, you are down to 100 microseconds," he added.

However, while Burton lauded the much lower latencies associated with Flash storage, he also acknowledged the capital outlay for the technology is much higher.

But, he argued, Flash offered much better value for money up to 50 times more dollar per IOPs than with a standard hard drive, justifying the initial expense.

Burton revealed the company will be releasing XtremIO, an all-Flash array that initially began shipping to selected customers in March 2013, to general availability before the end of 2013.

"This really is a specialised weapon," he said. "Because of its architecture, whether this array is empty or full, whether it's idle or busy, you're going to get sustained half-millisecond response times."

EMC acquired Flash storage specialist XtremIO in May 2012 for $430 million (266 million at the time), before it had started shipping any technology to customers. It appears the acquisition caused something of a lag in the technology's release as it had been expected to start shipping before the end of last year.

A further market trend identified by Burton is converged infrastructure.

"[If] the market for converged infrastructure...is not on your radar right now, my prediction [is that] in the next few years it will be," he told delegates.

"It is expected to grow 40 per cent...over the next four years in contrast to the two per cent growth for the individual components. So again, it is a trend that is on the up and we see these cloud building blocks' - server, storage and network - becoming increasingly deployed not just in IT departments but in service providers as well," Burton said.

Big Data was another of the event's key themes, with much of the discussion focusing on its use within companies, as well as practical questions around storage and analysis.

According to a survey carried out by EMC for the conference, 75 per cent of 455 IT decision makers said their organisation could be improved with better use of data and 37 per cent agreed their senior teams trust Big Data insights to make transformative business decisions.

A further 21 per cent claimed to have achieved a competitive advantage through the use of Big Data analytics, while 44 per cent said they thought the industries using these tools will be successful in the future.

However, 48 per cent said a lack of a clear business case or proven ROI was stopping them from adopting a Big Data strategy.

At a media roundtable at the Forum, Burton said this was likely because the business teams within their organisations do not understand what Big Data is.

"Until a business really internalises [what Big Data is] and understands the impact it can have and puts some people looking at what it can do, then this thing is like a hammer looking for a nail," said Burton.

"They've heard about it and they think it sounds good, but they don't know what they can do with it."

Burton added the 21 per cent who had already claimed to have gained a competitive advantage were the early adopters but the rest would follow in the coming years or be left behind.

Fusion-io is cementing its move downstream with the release of an SME-focused hybrid storage product that combines Flash technology with high capacity disk drives.

The firm's Fusion ioControl product is based on the technology the firm gained through its acquisition of hybrid storage appliance maker NexGen storage in April 2013.

Prior to the acquisition, NexGen already used the Flash-based storage vendor's ioMemory technology to create SME-focused hybrid storage appliances containing standard disk drives.

This, the company claimed, allowed NexGen to transform x86 server platforms into systems that boast the same kind of performance as an all-flash storage array.

At the time, the vendor also said the acquisition would help it make further inroads into the SMB market, despite Flash technology traditionally being viewed as the preserve of the enterprise because of its high cost.

Speaking to IT Pro, Chris McCall, vice president of the ioControl product line, said this is an issue the company has been able to side step because of the hybrid nature of the appliance.

Starting prices for the appliances begin at $55,000 for the 16TB, $88,000 for the 32TB and $108,000 for the 48TB arrays. "These are all starting prices, and users can scale from there," said McCall said.

He was also keen to stress the product is aimed at SMEs with at least 100 employees that already have some form of datacentre capacity they rely on to carry out "heavy duty IT", rather than one man bands.

Even so, the company has had to adapt the products to make them as easy as possible for SMEs to use, McCall explained.

"This market has very different characteristics and needs than the big scale-out server deployments in enterprise database acceleration [where Fusion-io's products are typically used]," McCall said.

"SMEs tend to think more around simple management, because there are only a few IT guys that have to manage the technology for everybody. They can't just focus on one app, which presents a number of management challenges."

Rather than use a Flash-based array to speed up the running of a particular application, iOControl is aimed at SMEs that might want to run large chunks of their infrastructure on the technology.

Particularly companies that need to ramp up the performance of their server or desktop virtualisation deployments.

The product is being released in UK and Europe this week through the firm's tranche of channel partners.

Buffalo Technology is strong in the consumer and SMB NAS appliance markets and its latest TeraStation 7000 series signals a move into enterprise territory. The family consists of three models with the flagship TeraStation 7120r Enterprise on review.

This 2U rack appliance employs a 3.4GHz Intel Xeon E3-1275 CPU partnered by 8GB of DDR3 memory. The latter can be upgraded but you'll need to replace the existing DIMMs as the motherboard only has two slots.

The Enterprise model is offered with all twelve hot-swap bays populated by 2TB WD Enterprise SATA II hard disks. You can't buy a diskless model and if a drive fails you must get your replacements from Buffalo.

You can't expand outside the box either as, unlike Synology's RS10613xs+ or Broadberry's CyberStore 316S-WSS, the 7120r doesn't support external disk shelves.

Buffalo also offers two standard 7120r models with 8TB or 12TB of storage and empty bays for adding extra drives later on. These are slightly different as they have a slower 3.1GHz Xeon E3-1225 CPU and 4GB of DDR3 memory.

Buffalo's NASNavigator 2 makes light work of initial deployment and share mapping

Deployment

Installation is swift as you use the NASNavigator 2 utility to auto-discover the appliance. Along with direct access to the web management console, it provides a handy network share browsing and drive mapping service.

The web interface is the same as that presented by Buffalo's desktop appliances and is easy enough to use. For file sharing, the appliance supports SMB, NFS and AFP and each protocol can be enabled or disabled with a single mouse click.

Access security to shares is good as the appliance supports a local user database or authentication via an NT domain or Active Directory. Share creation is simplified as all features are accessed from a single screen where you can decide on global read or read/write access, select which protocols are supported, assign a Recycle Bin and apply access permissions.

The 7120r supports all key network protocols which can be enabled and disabled with a single mouse click

Features

The 7120r supports share replication using Rsync but only to other TeraStations. It can manage scheduled backups of local data and when creating shares, you can enable the Backup option which allows them to be used as destinations.

The appliance has six USB2 ports for external storage devices but does not have any integral USB3 ports. If you want these you'll need to add Buffalo's optional USB3 PCI-e card which costs around an extra 35.

The bundled NovaBackup Business Essentials software makes light work of scheduled workstation backups and the price includes a 10-user license. Cloud backup features aren't as plentiful as those offered by Synology and Qnap as the 7120r only supports Amazon S3.

IP camera surveillance is another new feature although you'll need to use a Windows PC as a host for the Surveillance Server software. The appliance simply acts as a vault for camera recording as all schedules and camera configurations are managed at the host PC.

You must enable LVM for combined NAS and IP SAN operations but this action will delete any existing shares

IP SANs and LVM

Before using the appliance, think carefully about how you plan to deploy it. If you want to provide both NAS and IP SAN services you will need to reconfigure the 7120r first.

Buffalo implemented its LVM (logical volume manager) feature in all its latest NAS appliances to allow them to support IP SANs but there are caveats. Out-of-the-box, the 7120r Enterprise has all drives pre-configured as a single RAID-6 array with LVM disabled.

This supports a single logical volume which can be used for NAS shares or one big IP SAN volume. LVM allows multiple NAS and IP SAN volumes to coexist in the same array but enabling it will delete any pre-existing shares on that array.

Disabling LVM on the array will also delete any shares and IP SAN targets. However, with LVM enabled, you can create dedicated NAS volumes that can be expanded into free space on demand.

It's exceedingly well hidden but the 7120r does support thinly provisioned iSCSI virtual volumes

10-Gigabit performance

Along with the four embedded Gigabit ports, the appliance has two PCI-Express slots and supports a good range of single- and dual-port 10GbE adapters from Intel and Emulex. For testing we installed an Emulex OCE11102-NM dual port 10GbE adapter which was accepted without any problems.

We started performance tests over Gigabit and saw good results. Iometer reporting fast raw read and write speeds of 109MB/sec for a share mapped to a Dell R515 server running Windows Server 2012.

For 10GbE implementations, we recommend enabling Jumbo frames. Iometer reported a raw read speed of 682MB/sec for a mapped share without them which jumped to 847MB/sec with 9,000 byte frames configured.

IP SAN performance over 10GbE was even better with Iometer reporting a raw read rate of 880MB/sec for a thinly provisioned 500GB target. Buffalo warns that LVM may have performance overheads but during testing we could see no evidence of these. Conclusion

The TeraStation 7120r Enterprise scores highly for performance and also compares well on price with similarly equipped 12-bay rack appliances from Qnap. Support for optional 10-Gigabit networking is a valuable feature although the lack of external expansion counts against it and Buffalo's LVM feature isn't the most elegant means of implementing NAS and IP SAN support.

Verdict

Buffalo’s first enterprise network storage appliance is undeniably fast over Gigabit and 10-Gigabit. The TeraStation 7120r Enterprise is also competitively priced although expansion options are limited and its LVM feature requires careful thought about deployment.

Chassis: 2U rack

CPU: 3.4GHz Intel Xeon E3-1275

Memory: 8GB DDR3

Storage: 12 x 2TB WD Enterprise hot-swap SATA II hard disks

Array support: RAID 0, 1, 5, 6, 10, 50, 51, 60, 61, JBOD

Network: 4 x Gigabit

Other ports: 6 x USB2

Expansion: PCI-e X4, PCI-e X8

Power: 2 x 550W hot-plug PSUs

Management: Web browser

Software: Buffalo Surveillance Server (1 camera license), NovaBackup Business Essentials v13 (10 licenses)

Warranty: 3yr RTB

US-vendor Pure Storage has marked its launch in EMEA by pouring scorn on EMC's flash storage plans, while ruling itself out of becoming the flash market's next takeover target.

The company specialises in the manufacture of all-flash storage arrays, and claims it can offer the technology at a price comparable to what end users spend on mechanical disk-based storage.

Up to now, the firm has focused on building out its presence in the US, but recently opened an HQ over here and has set its sights on winning over UK enterprise customers.

Speaking to IT Pro, Scott Dietzen, CEO of Pure Storage, set out the firm's UK growth ambitions.

"Usually companies in this space are benchmarked [for growth] against NetApp and Data Domain, as each of them grew by an order of 40 per cent sequentially [from launch]," he claimed.

"We're out of the gate faster and growing faster than they did...and that's our expectation for every market that we enter."

The Flash market is an area that storage veteran EMC has made an effort to move into of late, through its acquisition of Xtrem-io last year and the subsequent release of its new Flash storage line.

However, Dietzen claims EMC is playing catch up with the rest of the flash storage market, and estimates his firm has an 18-month technology lead on its rival.

"EMC bought the start-up competitor closest to Pure in terms of architecture [Xtrem-io], but we believe we have an 18-month technology lead [on the]. They have a good team, though, so obviously the race is on," he said.

"We don't see anyone else in the market that's nearly as close to having the full recipe that we've launched with."

But that's not to say the company is resting on its laurels, he added, as there is still the possibility that another new entrant could emerge that leaves both firms for dust.

"It's entirely possible, but the hurdles to get into the market are on a par with the ones [vendors must overcome] to take on VMware, who've had a very nice run. Right now, though, we're starting to see other offerings that are [getting closer to VMware] technically," he said.

To mark Pure Storage's launch in EMEA, the company has taken the wraps off its "Love Your Storage" money-back guarantee programme, which will see the firm reimburse customers who expressed disappointment in the performance and reliability of its all-flash arrays, for example.

"We believe we have the best technology in the space and we will give you your money back if we don't exceed all of your expectations in that regard," Dietzen added.

He was also keen to stress the firm has no plans to follow Xtrem-io's lead and get itself acquired, claiming the company's growth to date means it should have no problem maintaining its status as an independent storage vendor.

"We've had numerous requests for M&A conversations. There was a bidding war for Xtrem-IO and all of the bidders talked to us before and after to express their interest, but we want to build an independent storage company," he explained.

"The rationale behind that is because we think we can do a better job serving customers as an independent than by joining a bigger company...who may put a block on us approaching customers where they have [long-standing] disk contracts."

EMC has confirmed that its XtremIO all-Flash storage array has started shipping to select customers ahead of its general release later this year.

The product has been in development for quite some time, and news of its release comes less than a year since the storage giant acquired Israel-based XtremIO for $430 million.

Before the deal was announced in May 2012, the flash memory-based storage start-up stated on its website that it was planning to ship its first products to customers in the second half of 2012.

The EMC-branded all-Flash storage array features a scale-out architecture that aims to provide a higher level of "functional IOPS" to applications that require high levels of random I/O performance.

Speaking to IT Pro, Richard Gordon, Flash business development manager for EMEA at EMC, said the product would be well-suited to VDI, database consolidation and server virtualisation projects.

"It's the answer for VDI [deployments] at scale, where the IOP requirement overtakes everything else," he said.

"In that type of environment, traditional SAN-based technologies can begin to struggle [which impacts on the user experience], which is why some larger VDI deployments have never made it to production."

EMC has also announced the release of its Xtrem family of Flash-optimised storage and server products, as well as the introduction of a new line of EMC XtremSF PCIe-based Flash cards.

The latter can be deployed as either direct attached storage (DAS) that sits within a server or coupled with EMC's server caching software to boost the performance of network storage arrays.

The products, coupled with the firm's existing Flash-friendly offerings, should put it ahead of the competition, claimed Gordon.

"If you look at the three quadrants I focus on: direct attached Flash, an all-Flash array, and extreme software cache, EMC is the only one with a presence in all three areas. That differentiates us straight away," he said.

"I can confirm is that our roadmap is being delivered upon in terms of capability and the rollout of the XtremIO platform. I'm really confident in our ability against our competitors."

However, Thomas Kejser, EMEA vice president and CTO at rival Flash storage player Fusion-io, claims EMC still has some way to go to beat its products on performance.

Speaking to IT Pro, he said his firm announced back in January that it had broken through the one billion IOPs barrier using eight HP ProLiant servers kitted out with eight of its Flash-based ioDrive2Duo drives.

"EMC announced that they went over a million IOPs [with its Xtrem SF Flash cards], and we see that as a complete non-event," he said.

"That's a thousand times lower than we announced earlier this year, so they've got a lot of catching up to do."

The worldwide storage software market grew by less than one per cent, according to a new IDC report.

The analyst company said that worldwide revenues increased by only 0.9 per cent over the past year. The storage software market closed the second quarter 2012 higher than compared to the same quarter last year, but only just.

EMC, IBM, and Symantec remained the top ranking storage software suppliers with 26.4 per cent, 14.7 per cent, and 14.6 per cent market shares, respectively. CommVault and EMC showed the largest organic growth during the quarter with 21.5 per cent and 7.4 per cent year over year increases.

The report found that data protection and recovery and archiving software markets were once again the two fastest growing segments with 2.4 per cent and 2.2 per cent year over year growth rates, respectively, or $1.16 billion and $404 million in total revenues.

Large enterprises contributed to the arrest in the market. Companies with more than 1,000 employees made up the biggest component of the market with $1.5 billion (926 million) in storage software investments during the quarter, but this represented a fall of one per cent compared to the same quarter a year ago.

Storage software investments made by small and medium companies countered this fall, managing to grow 5.2 per cent year over year to $574 million (354 million). Spending on storage software by governments and education also fell 1.5 per cent to $472 million (291 million).

Eric Sheppard, research director, Storage Software at IDC said the results of the second quarter were "mixed when viewed by supplier or functional market."

"Indeed, five of the top eight suppliers experienced revenue growth, but this was offset by declines within a few of the market's larger suppliers," he said.

"Most functional markets showed increased investments, but the increases were far smaller than the market had been experiencing over the past two years."

Sheppard added that these generally lower results could be partially attributed to suppressed economic growth in Europe, reduced government and education investments as well as "transitions specific to a few large suppliers."

Government departments spend an average of 236,000 each year on premium storage brands, according to the findings of a newly published Freedom of Information (FOI) Act request.

The request was made by open storage provider Nexenta Systems, and 44 out of 48 local and central government departments responded to it.

The results revealed these departments buy at least 101 TB of storage every year from big-name brands, such as IBM, NetApp, HDS and EMC.

It is incredible to think just how much of British taxpayers' money is being wasted on expensive storage solutions.

Typically, each terabyte costs them around 2,000, although some departments are spending as much as 5,000, the request showed.

The Government has endeavoured in the past few years to cut costs and create balance between open source and proprietary software vendors with varying results.

Evan Powell, the chief executive of Nexenta, said the public sector needed to radically rethink its storage buying strategy in light of continuing economic uncertainty.

"It is incredible to think just how much of British taxpayers' money is being wasted on expensive storage solutions due to a combination of vendor lock-in and general apathy," said Powell.

Powell said the government was unwilling to promote open storage because of "confusion, apathy and fear."

He also claimed that organisations, such as the US Army, Cisco, Korea Telecom and VMware had already embraced open source storage, netting them millions of dollars in savings.

Ryan Tyler, technical sales manager at storage provider VA Technologies, backed Powell's view.

"The public sector urgently needs help from IT partners who have the necessary skills and tools with open source storage to make the transition [from proprietary to open storage] as soon as possible," he said.

"The rapidly growing cost of maintaining and expanding on their legacy storage systems is proving too expensive, a fact that the government needs to quickly wake up to."

EMC has snapped up flash storage specialist XtremIO for $430m (266m), nearly a month after rumours linking the two firms first surfaced.

EMC appears to have been courting the flash-based storage vendor for a while, after its senior executives were reported to have visited XtremIO's offices in Israel last month to discuss a takeover.

XtremIO is the sixth vendor EMC has acquired in Israel, which is also home to one of its research and development centres.

In a statement, the storage giant said XtremIO's technology would fit in well with its existing range of Flash-based arrays, which are featured in its VMAX, VNX and Isilon ranges.

XtremIO was set up in 2009 and, according to a post on its website, was due to start shipping its first products to customers later this year.

In a statement, Pat Gelsinger, president and chief operating officer at EMC, said: "XtremIO brings an amazing technology with a fantastic team that's captured praise from early-view customers and the [IT] industry.

"We fully expect XtremIO technology, once introduced to market, to have a tremendous impact on our customer's ability to [make use of] the unique advantages of all-Flash storage across many of their most demanding applications."

EMC is in talks with XtremIO to purchase the company for between $400 and $450 million (248 - 279 million), according to reports.

Globes, a business news site from XtremIO's home country of Israel, published the rumour this morning, citing market sources who claim a deal is "close at hand."

Senior executives from the storage giant have reportedly visited the country for discussions, including chief executive (CEO) Joe Tucci and VMware chief technology officer (CTO) Stephen Herrod.

There are already links between the two firms as XtremIO's co-founder, Shuki Bruck, also set up Rainfinity, a file virtualisation platform he sold to EMC back in 2005.

EMC has been heavily investing in Israel for several years, acquiring five firms in the country and opening up a research and development centre there in 2011.

XtremIO was founded in 2009 and is known for its flash memory-based storage arrays. It claims "the power of flash" gives enterprise storage products better performance, whilst remaining simple, cost effective and eco-friendly.

EMC has been investing in flash technologies, with the likes of VFCache previously known as Project Lightning bringing a faster tier of storage into the server.

However, the Globe report says EMC is not the only firm courting XtremIO, with major rival NetApp reportedly making visits to the country as well.

The website did not mention the NetApp executives by name, but claimed visitors held the roles of head of business development and mergers and acquisitions, as well as the vice president for EMEA.

IT Pro contacted EMC for comment on the deal, but had not received a response at the time of publication.

EMC's hold on the storage market has led to a "bottleneck" in the industry many companies are finding it hard to break out of.

This was the belief of David Flynn, chief executive of Fusion IO, who met with IT Pro today to vent his frustrations.

He claimed NAND flash was a key component of storage, but the way EMC was trying to use it was too expensive and didn't offer enough performance.

"EMC are still so focused on their storage boxes," he said. "Yes, they can [upgrade them with flash] and offer you twice the performance, but they will charge four times as much."

By putting SSDs into the storage boxes EMC are trying to push an elephant through a straw.

There is no question in the storage industry that NAND Flash is the future for high performance, but the debate of where the technology should sit within the system rages between the two firms.

EMC has attempted to move it to the storage array, putting SSDs into its portfolio to speed up the read/write process. However, Fusion IO, has invented a new way of plugging flash straight into the memory.

By putting one of its ioDrives straight into the server, it removes barriers between the processor and traditional storage arrays, meaning the server can act as primary storage and access applications even faster.

"By putting SSDs into the storage boxes EMC are trying to push an elephant through a straw," said Flynn. "We are right next to the processor and don't rely on networking or [other hardware] performance."

He claimed the openness of the Fusion IO technology and its compatibility with other vendors meant end users only had to buy the one component, rather than hefty hardware in the back-end, which EMC is known for.

"For every $10 EMC costs, we can give you 1,000 times faster performance and cost just $1," added Flynn.

"EMC cannot afford to do it [Fusion IO's] way because they would lose thousands of dollars."

EMC do offer a caching technology for servers called VFCache, allowing a cached version of data to be stored before heading to the storage. However, the CEO said it didn't compete with the number of terabytes its own Fusion IO directCache could offer.

EMC still holds the number one spot in market share when it comes to storage. However, Fusion IO has the likes of Apple and Facebook on its customer lists, both of which is worked together with to get the product just right for their internal systems, and has Steve "Woz" Wozniak backing the firm as its chief scientist.

Flynn revealed there would be more products and market opportunities from his company next month just before EMC has its own annual conference but his company "didn't need to sell boxes" to give great storage capabilities.

"It is not about the hardware in the big boxes anymore," he concluded. "It's about elements of hardware working together with software to give the end user the best performance, as well as the best price."

Western Digital showed off its latest My Book Thunderbolt Duo external storage box at CeBIT this week.

Each box holds two hard drives for up to 6TB, and thanks to their use of the new Apple/Intel-developed Thunderbolt interface, up to seven can be daisy-chained off a single host, with data access at up to 10Gbps over copper cabling.

Among the others at CeBIT with Thunderbolt technology were MSI, Promise and OCZ.

MSI showed a PC motherboard with Thunderbolt fitted as standard. Until now, only a few PCs have had Thunderbolt as standard, most notably the latest Mac systems.

Promise's offer was its J4 disk array which takes 2.5-inch hard disks or SSDs (solid-state drives), and OCZ which said it would add the technology to its SSD range later this year.

Of course, while hard disks can use Thunderbolt, its high speed will be of bigger benefit to SSDs.

Also showing new storage at CeBIT was Qsan Technology, which unveiled two new unified storage devices for networks.

Top of the line was the U600Q, which offers two 10Gbit iSCSI ports and three Gigabit Ethernet ports, and is based around the Sun-developed ZFS network file system.

Qsan claimed that, thanks to ZFS and Qsan's own unified user authentication software, the U600Q can talk to pretty much any client, whether it runs Windows, Linux/Unix or MacOS, or connects via FTP or WebDAV.

TeamDrive Systems has added clients for Android and Apple iOS tablets and smartphones to its enterprise-grade secure filesharing platform.

The company announced the clients along with a new release of the service TeamDrive 3.0 at the CeBIT trade show in Hanover this week.

The service uses AES-256 encryption and its data centres are based in Germany, which has strict data protection laws.

The ability to use our own servers for shared cloud storage is fantastic.

The company said this means it can offer Dropbox-like ease of use, but without the same vulnerability to hackers or US federal snoopers.

There is flexibility too. Instead of hosting their encrypted files in the TeamDrive public cloud, customers can choose to use their own private server or another public cloud service, TeamDrive CEO Volker Oboda noted.

Furthermore, as TeamDrive's cloud is hosted within the EU, the service is fully compliant with the Safe Harbour rules.

Dr Anthony Beck, DART project champion and research fellow in the School of Computing at the University of Leeds, said that Safe Harbour is why his team uses TeamDrive rather than Dropbox.

"We can't use DropBox or other cloud data storage companies that are not a signatory to Safe Harbour," he said.

"Instead of our data being held on DropBox's servers in the cloud - DropBox actually uses Amazon's cloud service - we can host them on our own servers.

"The ability to use our own servers for shared cloud storage is fantastic. Essentially we have an unlimited storage capacity which we can easily share with colleagues throughout the world at any institution. This really helps collaboration. The smartphone application will provide the final piece allowing us to access data whilst 'on the go'."

Collaborative features of the TeamDrive mobile app include a photo-sharing tool and the ability to send RSA-2048-encrypted file sharing invitations.

The service can also be accessed from Windows, Mac and Linux PCs, and Oboda revealed BlackBerry and Windows Mobile apps are under development.

The new DS3612xs moves desktop NAS appliances into the big league as Synology's mighty cube is aimed at larger businesses with a need for speed and capacity.

It's the first to offer twelve hot-swap drive bays and also has a pair of high-speed Infiniband ports that each accepts a DX1211 12-bay expansion unit to boost raw storage to up to 108TB.

There's much more as the DS3612xs also sports an Intel Core i3 CPU and its base 2GB of server grade ECC DDR3 memory can be upgraded to 6GB. Along with quad Gigabit ports, the appliance also has a PCI-e slot that accepts single and dual-port Intel or Emulex 10-Gigabit (10GbE) cards.

The DS3612xs is very well built and a single keyhole above the drive bays allows all to be locked. At the back are the two Infiniband ports and four USB2 ports for adding other external storage devices or printers for network sharing.

The DS3612xs has two redundant cooling fans plus a very large passive CPU heatsink, making it extremely quiet.

USB3 and eSATA ports are conspicuous by their absence, and the DS3612xs only supports SATA II speeds. SATA III drives can be used, but they'll only run at 3Gbit/s although our performance tests using SATA II drives suggest that this isn't an issue.

We fitted an Emulex dual-port 10GbE card prior to testing, which was easily done as the DS3612xs' top panel can be removed separately to access its PCI-e slot. Memory upgrades aren't so easy though, as the motherboard must be completely removed from the side to get at the DIMM slots.

Incidentally, the Gen 2 8X PCI-e slot in the DS3612xs had no problems working with the Emulex card. In a recent review of the Thecus N8900 appliance in our sister title PC Pro, however, we found that its 4X PCI-e slot doesn't support Emulex 10GbE cards.

The appliance can be upgraded to 10-Gigabit as we had no problems with our Emulex dual-port card.

Installation is handled smoothly by the Assistant utility and the latest Synology DiskStation Manager (DSM) software can also be loaded during this process. The appliance's web interface is also well designed, and is easily be customised using drag and drop operations.

RAID options are extensive and the DS3612xs' Disk Group feature supports multiple volumes in a single RAID array. Odd-size drives in search of a good home can also be used, thanks to the Hybrid RAID option that creates arrays regardless of drive make and capacity.

IP SAN support is also good as the DS3612xs can manage up to 64 iSCSI targets with thin provisioning. Select the latter option during target creation and will only use a small amount of drive space that will have blocks added dynamically as space is used up.

Synology offers a heap of extra features as standard, including FTP, mail, web, photo and multimedia servers, plus a Surveillance Station with some good IP camera recording and motion-detection features. Backup features range from locally managed jobs on the appliance, NAS-to-NAS replication, Rysnc support and the basic (but useful) Data Replicator 3 workstation backup software.

The current version of Synology's DSM provides a slickly designed interface that's very easy to use.

Real-world speeds for read/write are impressive and backup operations are similarly quick.We created a RAID-5 array using four 2TB Seagate SATA II drives for our performance tests and linked the appliance over 10GbE to Dell PowerEdge R515 and Broadberry dual Opteron 4100 rack servers. Both were running Windows Server 2008 R2 and equipped with Emulex 10GbE dual-port adapters.

With a share mapped to the Dell server, Iometer reported an impressive raw read speed of 670MB/sec. On a second share mapped to the Broadberry server, Iometer reported a cumulative read speed of 1143MB/sec.

Real world speeds are just as impressive. Copying a 2.52GB video file returned read and write speeds of 290MB/sec and 280MB/sec, respectively; while FTP speeds with FileZilla stepped up to 344MB/sec read and 303MB/sec write. Backup operations are similarly quick and a 22.4GB collection of 10,600 files copied to the appliance at 116MB/sec.

The latest Synology DiskStation Manager 4.0 beta adds a wealth of new features, including an improved user interface that adds seven new widgets for monitoring system health and resources, connections, backup jobs and logs. A free anti-virus package is also available and the already excellent Surveillance Station gets even more features, including iSCSI LUN backup.

The latest DSM 4.0 beta offers an improved user interface with extra widgets for monitoring the appliance.

Synology has also added cloud file sharing and backup services to the DS3612xs. The ezCloud service can use Synology's free DDNS service for secure remote access to shares, while its Cloud Station tool provides a simple, but effective, syncing service for up 32 file versions.

Cloud Station keeps multiple file versions on the appliance, which can be searched through for restoration via the DSM 4.0 interface.

The Cloud Station package need to be installed on the appliance and then registered. Synology's QuickConnect feature does away with the need to create router port forwarding rules though, and instead a code just needs to be entered in the Cloud Station interface.

Synology's QuickConnect feature does away with the need to create router port forwarding rules.

Once a local folder has been selected, Cloud Service will then sync it with a predefined folder on the appliance in the user's home directory whenever its contents change. The system worked very well during out tests, although we did find that Norton AntiVirus Sonar will try to block the sync service.

Not only is the DS3612xs currently the largest desktop appliance on the market, but it's also one of the fastest. Its enormous expansion potential makes it a good long-term investment and the number of features Synology is packing into its NAS appliances makes many competing products look positively archaic in comparison.

Verdict

This high capacity desktop appliance delivers top performance over Gigabit and 10-Gigabit networks, and offers a huge expansion potential via high-speed Infiniband ports. Synology’s list of storage features is probably the best currently available and the next version of its DSM will add useful cloud backup services.

Chassis: Desktop CPU: 3.1GHz Core i3-2100 Memory: 2GB DDR3 ECC expandable to 6GB Storage: 12 x 3.5in.or 2.5in SATA II hard disk hot-swap bays RAID: Software managed Array support: RAID0, 1, 5, 6, Hybrid, hot-spare and JBODs Network: 4 x Gigabit Other ports: 4 x USB2, 2 x Infiniband expansion Expansion: 1 x PCI-e X8 Gen 2 Management: Web browser Software: Synology Assistant, Data Replicator 3 Options: Diskless DX1211 12-bay expansion unit, £750 ex VAT

Dell has acquired app backup provider AppAssure as it looks to boost its already flourishing storage business.

Virginia-based AppAssure claims to offer "near-zero recovery time" for apps on virtual or physical infrastructure, or hosted in the cloud.

The company, whose 230 employees look set to be retained by Dell, believes it has the "industry's fastest backup and recovery of virtual servers" too, covering VMware, Hyper-V and XenServer hypervisor environments.

AppAssure enables Dell customers to seamlessly move and replicate data across our existing platforms.

"AppAssure's unique architecture delivers innovative cloud-enabled backup and replication solutions that meet the challenges of protecting the explosive growth of data in both virtual and physical environments," said Brad Anderson, presiden for Dell's Enterprise Solutions Group.

"At a practical level, AppAssure enables Dell customers to seamlessly move and replicate data across our existing platforms - from an EqualLogic array in a remote office to a Compellent array at a data recovery site."

Dell said it would invest in additional engineering and sales capability to grow the AppAssure business.

Financial details were not disclosed.

The acquisition marks another major play from Dell in the storage space, coming after its Compellent purchase last year which has helped the company post solid results from that side of its business.

Dell recently ditched EMC as a major partner as the Texan company looks to follow its own storage path.

Dell's quarterly results released earlier this week showed the company's storage division was the fastest growing part of its business.

In the fourth quarter, Dell saw its storage division grow 33 per cent, whilst Compellent had sequential growth of over 60 per cent.

Rival HP, which had a bidding war with Dell over 3PAR in 2010, saw sales of enterprise servers, storage and networking equipment drop 10 per cent over the last quarter.

"Dell's acquisition of AppAssure is a strategic industry move given its relentless focus on providing innovative solutions to protect customers' applications and data in cloud, virtual and physical environments," added Najaf Husain, president and CEO of AppAssure.

ioSafe this week unleashed its product family onto the UK market, giving us storage it claimed could survive any disaster.

The US firm targets its products at small and medium businesses (SMBs) or branch offices for larger firms, although some solutions fall into the consumer category.

It sells the devices through channel partners and has just signed an agreement with Micro-P to get distribution moving in the UK.

The idea behind the range is that it can withstand fire, flood or even theft

"Now is the right time to expand into the UK," Robb Moore, chief executive (CEO) of ioSafe, told IT Pro. "We have had lots of requests for our products over here and the UK is a fantastic marketplace."

"If you combine the UK's understanding of data storage and its care about security, it is the right move for us."

The idea behind the range is that it can withstand fire, flood or even theft thanks to built in physical protections to the hard drives and SSD-based technologies.

"Encryption and network security are often used to protect data," read a statement from ioSafe. "Wildfires and hurricanes care little about whether the data is encrypted [and] firefighters won't ask your permission before hosing down your entire computer system."

"Physical security of data is often overlooked. Nearly every business, large or small has vulnerabilities [and] ioSafe specialises in hardware to help protect vulnerable data wherever it may reside."

There are two main lines of storage products from ioSafe, the desktop range and the portable portfolio.

The Solo and SoloPro offerings come with either HDD or SSD and are both bulky, designed to be kept in-house. However, with this bulk comes resilience, as the company claims they can withstand 30 minutes of 1,550F heat or up to 72 hours under 10ft of water without losing data.

The Rugged Portable drives are more about the risks data faces on the move. Again with the choice of HDD or SSD, these products can survive the crush of up to 5,000lbs and a drop of 20ft whilst still retaining their data, claims ioSafe.

For more reassurance, ioSafe offers a one-year warranty alongside a one year data recovery service, which can be extended up to five years if paid for. The promise with these services is to get replacement hardware after a disaster occurs with all your original data restored to it.

"We have a 99.99 per cent success rate of keeping and retaining that data so it makes a big difference to businesses and people," added Moore.

"We have products shipping now and, although we are just getting started, we are really excited about expanding here in the UK."

SMART Global Holdings today announced the formation of a new independent company, built to focus on flash and solid state disks (SSDs) for enterprises.

SMART Storage Systems plans on taking advantage of the growth in the SSD market as more companies begin to move from traditional hard drives to the speedier storage technology.

2012 is the year of SSD it is finally becoming a proven platform for high-speed transactional storage.

As a standalone organisation, it will have its own executive team separate from its parent company, as well as independent engineers, sales and marketing employees, and a dedicated manufacturing plant in Malaysia.

"Organisations in nearly every market sector are finding that the need to access and process information is outpacing the performance capabilities in traditional hard-drive based server environments," said John Scaramuzzo, president of SMART Storage.

"SMART Storage is not only focused on overcoming these bottlenecks by bringing high-performance SSDs to market, but also on solving the inherent endurance issues of flash."

The year of SSD' has been touted for sometime now, and is a belief held by SMART Storage and SMART Global Holdings. Now, it seems, industry analysts are on-board as well.

"Yes, 2012 is the year of SSD it is finally becoming a proven platform for high-speed transactional storage, and is now available as a tier 0 option in the vast majority of new storage systems," Clive Longbottom, founder and senior analyst at Quocirca, told IT Pro.

"As a tier -1 solution, companies such as TI and Fusion io are also pushing the envelope with their PCI-X SSD systems that push performance even further."

However, he did question the idea of having a purely SSD focused company for the storage firm, as although pure SSD offerings will be useful for OEMs and for certain applications, most customers will be looking at a hybrid approach.

"The long term successful storage vendors from any area will be those who can figure out how to build in such intelligence as close to the storage components as possible and those who have a balanced portfolio will be better positioned for this than any company attempting to specialise in just one type of storage," concluded Longbottom.

SMART Storage has been created from the funding acquired by the parent company in August last year from Silver Lake Partners and Silver Lake Sumeru.

Dell has fallen short of revenue expectations in its fourth quarter, despite claiming some big gains in its business solutions and services divisions.

Revenue was up by two per cent for the company's fiscal fourth quarter 2012 to hit $16 billion, although that did not meet Wall Street predictions.

Compared to the fourth quarter in 2011, net profit slipped 18 per cent, from $927 million to $764 million.

I don't see (smartphones) as a top priority for us, either in the consumer or the overall business.

The company has shifted strategy over recent years, looking to become a bigger services player as it faces up to dwindling PC demand.

Financials had been improving, with solid results in February and August last year as the company sought to establish the "new Dell."

Yet fortune has not favoured Dell, with the Thai floods hitting its hardware supply chain as hard drive manufacturers in south east Asia struggle to get back on their feet.

The turbulent economic climate has also caused Michael Dell and co trouble, as public sector spending has shrunk.

Revenue from public bodies decline one per cent, whilst consumer revenue dropped two per cent.

President for Dell in EMEA Aongus Hegarty told IT Pro the company still considered the consumer sphere important, although it was not as much of a focus as enterprise.

In particular, Hegarty said smartphones would not be a major priority at Dell.

"I don't see [smartphones] as a top priority for us, either in the consumer or the overall business. Our greater priorities are in the PC on the client side and notebooks, and tablets later this year," Hegarty said.

"With smartphones, we will continue to have products in that area so that we can provide the full range of devices.

"Our priority focus is around our enterprise business in solutions and services, but consumer is also important. As we see a trend of more consumerisation into commercial businesses we think it is quite important that we have a strong healthy business there that we invest and develop."

Hegarty also pointed to Ultrabooks as a key market for Dell in the future as it looks to reassert itself in the PC market, where it has lost ground to the likes of Lenovo in recent months.

Highlights

Notable successes for Dell came from its datacentre products as it looks to provide for the whole stack, with storage proving the biggest hit.

Since Dell bought Compellent, after missing out on 3PAR to HP, storage revenues have shot up. The company also ended its storage partnership with EMC last year as it sought to bolster its own IP.

In the fourth quarter, Dell saw its storage division grow 33 per cent, whilst Compellent had sequential growth of over 60 per cent.

"This is, clearly for us, highlighting clear customer needs in those areas and we're pleased we've invested in Compellent, Equallogic and Dell IP in these areas," Hegarty added.

He promised announcements in storage and servers in the next quarter, but would not be drawn into talking about any flash strategy in either of those two departments. The company has talked up flash, but has not made as big a splash as companies such as EMC have.

The company's transition to more of a services-oriented beast is paying off, even if results have been hampered by dabbling in the consumer space.

Dell Services grew to represent 14 per cent of Dell's business in the fourth quarter, whilst server and networking revenue grew six per cent.

As for the next quarter, Dell expects revenue to decline approximately seven per cent sequentially, "which aligns with normal seasonality adjusted for the fourteenth week," it claimed.

Greenbytes today announced its latest storage array, hoping to bring deduplication technology to the masses.

Despite being aimed at the small and medium business (SMB) market, the Solidarity' box has high performance levels.

The storage array is based on solid state technology and contains capacities of between 3.5TB and 13.5TB. The company claims performance of 120,000 4K IOPS and the appliance features dual controllers, able to operate independently, with four 1 GbE and two 10 GbE network ports.

However, it is the inclusion of the deduplication technologies that is getting the industry excited.

Innovations are typically first adopted by larger businesses and slowly trickle down to smaller companies.

Deduplication has been a staple of enterprise storage for some time, with Data Domain - now owned by EMC - leading the way in pioneering the technology. However, the technology has remained external from the SMB market.

"Innovations are typically first adopted by larger businesses and slowly trickle down to smaller companies as they become more affordable and user-friendly," Simon Robinson, research director at The 451 Group, told IT Pro.

"But, [its move to the SMB market] also speaks to the fact that smaller organisations are dealing with the same data growth challenges as larger businesses, and have even fewer skills/resources to deal with it, especially as they are looking to get to grips with server and desktop virtualisation - which can create large amounts of duplicate data."

With GreenBytes especially though, Robinson felt the use of SSD technology also necessitated deduplication technology.

"Storage optimisation and efficiency technologies, such as deduplication, are increasingly necessary components of storage systems that utilise solid state," he said.

"This is still an expensive resource, so squeezing the most value out of it is a paramount concern for cash-strapped IT departments, large and small."

GreenBytes offers inline deduplication which it claims reduces the storage capacity needs by a ration of 12:1 for virtual instances and up to 8:1 for databases. It also provides global deduplication for compressed replicated data.

Michael Robinson, vice president of marketing communications at GreenBytes, told IT Pro it was also about simplicty, helping those SMBs which analyst firm the 451 Group said had "fewer skills/resources" to look after the datacentre.

"For the SMB, Solidarity means the end of complex, HDD-centric, multi-tier storage that can't keep up with I/O and throughput requirements of today's mixed SME storage missions," he said.

"GreenBytes' single tier hyper-performance arrays are priced as solutions for SME storage missions from virtualisation through replicated DR."

The GreenBytes Solidarity storage array will begin shipping in the first quarter of 2012.

After dallying in more consumer spheres than perhaps it should have, HP is driving forward a strategy focusing heavily on IT infrastructure.

Meg Whitman is placing greater emphasis on areas such as storage, rather than getting distracted by the likes of smartphones and tablets.

With this in mind, we spoke to the head of the storage division across Europe, the Middle East and Africa (EMEA), Chris Johnson, to find out how the partner conference has helped him and what we can expect from HP storage this year.

How has the show been for you?

The show has been great. We obviously had the customer and partner show in Vienna in November which was great and here is the same, just with the partners.

You get extremely good dialogue, you touch everybody, get a great community together in one big hit. We call it speed dating and it works.

But, with this event being just partners, is it a tougher crowd to please?

We are a community together. It is different from customers. There you are trying to build a relationship and build a proposition, whereas HP has a very loyal partner community and it is really that synchronisation of friends together.

You get extremely good dialogue, you touch everybody.

We are here to understand what they need, make them money... it is a very natural relationship, much more so than other companies I have worked at.

HP just has that scale of channel, that history which runs into decades sometimes. It means getting together to work out how we can be successful together is fun, as well as being really productive.

The major launch this week was in servers and with Project Voyager. How does that affect what you are doing in storage?

Well, our strategy is converged infrastructure, so where we benefit in the market place is that we have market leadership in storage, servers and networking and we engineer that together.

We are striving to be technology leaders, not followers... we are trying to push the boundaries as a market leader should. In storage, for the first time in a long time, we have also got the capability to do that.

Converged Infrastructure from HP is split into products for traditional datacentres, virtual datacentres and then the cloud. Is the cloud where the growth in sales has been?

It is an interesting space in pure storage terms. I mean, 3PAR forms part of the HP cloud generic strategy. In VirtualSystem terms - HP's cloud in a box offering - this is a very interesting part of the market.

Other vendors with converged system offerings, typically multi vendor... are focusing on the high end of the market where customers have big teams of server, storage and networking people.

VirtualSystem, we believe, is something that is going to become very strong in the channel for the mid-market. [It's for] those who don't have the time, staff and skills to go through the virtualisation journey and want to go straight to the full cloud system.

Channel partners have all of the skills to sell each part, now they have something new to sell that comes in one box and differentiates from the market.

It is clear the message is our core business is infrastructure. The good news is we have developed storage, servers and networking to be disruptive.

I bumped into Dave Donatelli [executive vice president and general manager of enteprise storage, servers and netwokring at HP] in the corridor earlier and he is delighted. He did a great job of getting us where we are.

You can feel the ground swell with excitement of being back to HP core values.

What Meg [Whitman] adds to this is that focus back to the core values of HP, that we are an infrastructure company. Yes we have the software, yes we have the services and some great development around that... but we are a infrastructure company at our core.

It must be nice for the infrastructure guys to be centre of attention again after all these years?

Well, I am not a long term HP guy, but even in the nine months you can feel the ground swell with excitement of being back to HP core values. It feels great.

R&D-wise, are you guys looking to develop in-house more for storage in 2012?

There are going to be big things for 2012. I can't be too specific but we have a fantastic storage engineering team.

If you imagine the evolution from the traditional portfolio to the acquisitions coming in, there is a lot they are getting to grips with. But, some real groundbreaking stuff is coming, not just enhancements to this portfolio, but quite a lot of exciting stuff to get our teeth into.

Will the development still focus on cloud as it has in the past year?

I think we aim to be specialised in networking, servers and storage and deeply specialised because customers expect that, but it has to be integrated. That is what you are going to get from HP and that is why it feels exciting because I don't think a lot of other vendors get to do that. Expect some more from us in the next six months.

The focus on HP Labs is indeed exciting. Does that mean you don't need to go out and buy up any more companies?

Absolutely. We are engineering [what we have]. We are not really missing anything you can go out to buy.

What we have we will evolve and develop and that is the exciting bit. If you Google HP Labs, you will see a bunch of interesting stuff.

It will be a frenzied year in 2012. Expect a lot and hopefully when we next speak I can report a lot of growth.

Google is close to launching a cloud storage service, according to people familiar with the matter, moving it into closer competition with the likes of Apple's iCloud and Dropbox.

Google is to call its storage service Derive and will allow people to store their files in the cloud and retrieve them from whatever device they are on, the sources told the Wall Street Journal.

Collaboration features will also be added in to the service, the source said, although they were vague about the actual launch date, saying it could be weeks or months until the product arrives.

It will be free for most businesses and consumers, although as with Google's other cloud offerings, a fee will be charged for those who want to use up large amounts of storage in the company's data centres.

A Google spokesperson declined to comment.

Google already has a significant number of cloud products letting people store files and then get them on whatever device they are using Google Apps being a prime example. The company also launched its Google Music service last year.

However, it does not have one place in which to store various kinds of data.

The company was working on plans for a cloud storage service some years back. Now CEO Larry Page led the initiative, dubbed G Drive, which was due to launch in 2007 but never made it to market.

Earlier this week, F-Secure announced its own cloud storage offering. However, the product is being marketed for operators to use, meaning the likes of Virgin could soon be pushing out some fresh iCloud rivals.

Researchers today revealed a new technique of recording data magnetically, which they claimed would make the process "hundreds of times faster" than current hard drives.

The four year study was led by the University of York's physics department, partnering with Radboud University in Nijmegen, and involved a close working relationship with Seagate.

It claimed recording data by using heat would improve both speed and energy efficiency when it comes to writing data onto storage.

For centuries it has been believed that heat can only destroy the magnetic order.

Today's storage technology uses the standard magnetic theory of North and South poles of a magnet being attracted to each other, with the same directional poles repelling one another. Then, by introducing a third magnetic field, the poles are inverted and write a piece of data to the storage.

By increasing the strength of the third magnet, the speed of the information being written could be ramped up.

The new research shows, rather than adding a third magnetic field, heat could be added in the form of a ultrashort heat pulse' and have a much stronger reaction, speeding up data recording hundreds of times over.

"For centuries it has been believed that heat can only destroy the magnetic order," said Dr Alexey Kimel, from the Institute of Molecules and Materials at Radboud University, who also took part in the research.

"Now we have successfully demonstrated that it can, in fact, be a sufficient stimulus for recording information on a magnetic medium."

Professor Roy Chantrell, from the department of physics at York, told IT Pro a new project to extend the research had already received funding and will start in the next few months, but it will be some time until the technologies hit the market.

"The next phase of magnetic recording technology will be Heat Assisted Magnetic Recording, where a laser is used to heat the medium to allow recording to take place," he said.

"This will span up to maybe 10 years, after which new technologies, such as our proposed approach, will be needed."

The research has been published in February's Nature Communications journal.

EMC today launched VFCache, a flash caching module for servers announced last year under the Project Lightning moniker.

It marks another foray away from the storage array for EMC as it looks to bring its flash expertise to the server space.

Testing showed a three times throughput improvement and a 60 per cent dip in latency, the storage giant claimed.

If that means bringing storage to the compute makes sense, that's what we'll do. And that's kind of what we're doing with VFCache.

"This is a product that extends EMC's FAST technology into the server. It is a combination of hardware and software, the hardware being a PCIe card, which is a very exciting technology because if really changes the performance landscape it's even faster than flash inside a storage array," Barry Ader, EMC's senior director of storage product marketing, told IT Pro.

"We marry the hardware with some software intelligence that sits inside of the server, the operating environment, to determine exactly what data is hot and make sure we keep the hottest data in the server, closer to the application, closer to the CPU complex."

Data held in the cache is still pushed out to the storage array for added security, Ader said.

"This is a very nice complement to the VMAX and VMX storage devices," he added.

"VFCache will work with a variety of servers that are out there. We've qualified this with IBM, Dell, HP and our Cisco UCS partners. We have qualified VFCache with the majority of server vendors that are out there."

VFCache will either be sold to IT departments directly via EMC to line up with the vendor's storage products, or through channel partners, including server and storage companies.

The main competition for EMC in this area will be Fusion-io, which has been pushing out PCIe cards to the industry a few years ago.

"The difference here is that they [Fusion-io] went after the early adopters," Ader asaid. "They sold these cards as DAS [direct-attached storage] cards rather than cache devices. So that caused a lot of problems for their customers.

"Really mainstream customers weren't able to leverage this it didn't solve their scalability, shareability protection problems."

VFCache will also be getting deduplication technology added to it in the coming months.

Thunder coming

EMC also revealed plans for an entirely separate appliance, which will essentially do the same job as VFCache but considerably quicker.

"It's for those high-end use cases. A good example might be a high-end trading application... it might not be for the mainstream."

The first Thunder devices will be trialled with early adopters in the second quarter. Wider adoption will be opened up later in the year

EMC in server space?

Ader would not be drawn into saying whether this foray into the server space would mark EMC's eventual move into actually manufacturing servers.

"As our customers are looking for different ways to manage their information, of course EMC is going to look at these different ways that we can leverage technology like flash disrupting types of enabling technology.

"There are ways for EMC to enhance value to customers by not just going into the server space. That's not the intention. The intention here is, there are different ways to create value. In some cases that might mean bringing storage to the server.

"If that means bringing storage to the compute makes sense, that's what we'll do. And that's kind of what we're doing with VFCache."

Companies such as Cisco and HP are trying to pack more into servers, whether with networking or storage capabilities.

Dell, meanwhile, recently broke off its storage relationship with EMC, hinting it would be dallying in smarter servers with potentially greater storage integration.

Scientists at IBM and the German Center for Free-Electron Laser Science have built what they claim to be the world's smallest magnetic data storage unit.

According to the scientists, their technology enables storage more than 100 times denser than current hard disks and solid-state drives, and could lead to significantly smaller drives for mobile data in future devices.

As components continue to shrink, the march continues to the inevitable end point: the atom.

The scientists said they had successfully stored one bit of data in only 12 magnetic atoms, and claimed the ability to manipulate matter by its most basic components atom by atom could be vital as conventional miniaturisation becomes unsustainable.

"The chip industry will continue its pursuit of incremental scaling in semiconductor technology but, as components continue to shrink, the march continues to the inevitable end point: the atom," said Andreas Heinrich, lead investigator into atomic storage at IBM Research.

"We're taking the opposite approach and starting with the smallest unit - single atoms - to build computing devices one atom at a time."

IBM said nanostructures built to exploit an unconventional form of magnetism called antiferromagnetism could allow people and businesses to store 100 times more information in the same space.

The team engineered 12 antiferromagnetically coupled atoms that were able to store information for "a few hours" by taking advantage of their inherent alternating magnetic spin directions.

This property, the scientists said, enabled them to pack adjacent magnetic bits much closer together than was previously possible without disrupting the state of neighbouring bits.

Costs for Fibre Channel (FC) and IP SANs have dropped significantly over the past year, but many SMBs still find them too expensive and prefer to use direct attached storage (DAS) to expand server capacity. Infortrend has traditionally focused on delivering affordable business storage products and its latest EonStor DAS appliances are designed to share the storage wealth with multiple application servers.

On review is the ESDS S16S-R2240-4 which is a 3U system that can accommodate up to sixteen hot-swap LFF hard disks. The price we've shown is for a diskless appliance, but for testing we were provided with a full complement of high performance 600GB 15,000rpm 6Gbit/s SAS disks. Infortrend also offers 10,000rpm SFF hard disks and lower cost 1TB, 2TB and 3TB 7,200rpm nearline SAS drives.

Hardware resilience is good with the review system including a pair of redundant, hot-swap controllers. Each one has four 6Gbit/s SAS host ports so you can link four servers using dual-redundant links for high availability (HA) or eight servers in a non-HA environment.

Each controller is equipped with 2GB cache memory which can be upgraded to 4GB. They each have a battery backup as well as Infortrend's CacheSafe feature which writes all cached data to flash memory for extra protection in the event of a prolonged blackout.

Expansion is very good as each controller has a dedicated 6Gbit/s SAS port for adding either J2000-S or J2000-R expansion shelves, allowing for a maximum of 112 disks. Both are very similar 3U units with dual power supplies and fan modules, but the R' model has four SAS ports allowing the use of multiple fault tolerant paths to the main controller unit.

Each controller has a dedicated network port for web management, although the RAIDWatch browser interface isn't that easy to use and yet is too basic for our liking. Once you've assigned IP addresses to the controllers, it's far better to install the bundled SANWatch application suite which is more sophisticated.

However, it's worth taking a moment to get to grips with how storage is handled by the appliance. Logical drives are collections of physical hard disks grouped together in a RAID array. Once you've created a logical drive you can then create logical volumes. A logical volume can be confined to a single logical drive or it can be spread over multiple logical drives for even greater fault tolerance.

The SANWatch Commander provides a complete overview of all Infortrend storage appliances present on the network.

A logical volume comprises the total capacity of all included logical drives. Volumes can be divided into partitions. These are the actual storage devices that will be presented to your servers and are mapped to either controller and assigned a unique LUN.

Plenty of backup options are available. Snapshots provides point-in-time copies of partitions which can be used for rollback and recovery purposes. Source partitions can also be copied or mirrored to target partitions on other volumes. Once a mirror has completed, the pair can be split and the target mapped to a new LUN.

This is useful for offloading backup tasks as the target can be removable media attached to another server, reducing any potential performance lag on the source. Once the backup has completed, the source and targets are resynchronised and then split again ready for the next backup.

Optional thin provisioning allows you to create partitions that only use a small amount of space, but appear much larger to the host servers. As capacity is used, the appliance dynamically allocates blocks to the partition and when all the space is used up you can expand a partition into any free space on the logical volume.

Using the Iometer utility configured with four disk workers, a 64KB transfer request size and 100 outstanding I/Os we saw the Dell server return impressive raw read and write throughputs of 800MB/s and 685MB/s respectively.

A handy performance monitor provides a real time view of read and write activity for each physical hard disk.

With the Broadberry server connected to the same controller and using the second partition we saw cumulative read and write speeds of 1308MB/s and 712MB/s. This showed some contention for controller resources was occurring, although these speeds are impressive nonetheless.

For real world performance we saw drag and drop copies of a 2.52GB video clip to and from the Dell server return speeds of over 400MB/s. Our collection of small files was also handled well as a 22.4GB folder of 10,500 files was copied from the partition to the server at a rate of 160MB/s.

The EonStor S16S-R2240-4 scores very well for performance as only expensive 8Gbit/s FC SAN appliances will be able to match these speeds. It's also good value, although a close competitor is HP's StorageWorks P2000 G3 appliance.This is very similarly priced and, amongst its many connection options, supports 6Gbit/s direct attach SAS. Expansion and resilience are just as good, although the EonStor's larger 3U chassis does provide a higher base capacity than the StorageWorks's 2U head unit making it better value.

So what's our verdict?

Verdict

The EonStor acquitted itself well in our lab tests showing performance equal to more costly 8Gbit/s FC SAN appliances. It’s easy to deploy, provides excellent fault tolerance and expansion potential and it's good value too.

Chassis: 3U rack Storage: 16 x SAS drive bays Power: 2 x 530W hot-plug supplies Two hot-swap RAID controllers each with the following specification: Memory: 2GB cache (expandable to 4GB) with battery backup and CacheSafe Array support: RAID 0, 1, 3, 5, 6, 10, 30, 50, 60, JBOD, global hot-spare Host interfaces: 4 x 6Gbit/s SAS ports Network: Gigabit Ethernet management port Expansion: 6Gbit/s SAS port Management: Web browser, Infortrend SANWatch Suite

HP announced plans to stuff ARM-designed chips into its servers, as it looks to create super energy-efficient datacentres for cloud providers and large enterprises.

It marks a significant moment for ARM, which has done little to challenge the dominance of the x86 architecture used in the majority of servers today.

It also forms part of Project Moonshot a multi-year HP initiative designed to support those moving to "hyperscale environments."

HP Redstone is designed for testing and proof of concept. It incorporates more than 2,800 servers in a single rack.

HP is essentially looking to help companies truly consolidate their data centres, thereby making them more eco-friendly and efficient.

The company introduced the HP Redstone Server Development Platform, which will feature Calxeda's EnergyCore ARM Cortex processors. The system-on-a-chip processors should help companies basing deployments on the Redstone platform use 94 per cent less space in their data centres, HP said.

ARM launched its energy-efficient Cortex-A7 MPCore chip design last month with much fanfare around their power consumption capabilities. The British chip developer claimed its new processor design was five times more energy efficient than its predecessor, while also delivering significantly greater performance.

ARM won't be the only Redstone partner, however. HP confirmed future Redstone server deployments would come packed with chips from Intel and "others."

"HP Redstone is designed for testing and proof of concept. It incorporates more than 2,800 servers in a single rack, reducing cabling, switching and the need for peripheral devices, and delivering a 97 per cent reduction in complexity," HP said.

Select customers will be able to get their hands on Redstone offerings from the first half of next year.

To supplement the Redstone platform, the US tech giant launched the HP Discovery Lab. This services offering will let customers test out Redstone deployments before making any purchase. The first lab will open in Houston in January, with European facilities expected soon.

"The volume of data processed in financial markets has increased exponentially, and traditional scale-up or scale-out architectures are struggling to keep up with demand without vastly increasing cost and power usage," said Niall Dalton, director of high-frequency trading at Cantor Fitzgerald, a company looking into the Redstone technology.

"HP is taking a holistic approach to solving this problem and working to bring unprecedented energy and cost savings for tomorrow's large-scale, data-intensive applications."

At its foundation, the DX90 S2 uses a 2U head unit with single or dual RAID controllers with the latter run in active/active mode. The hard disk backplanes are now connected over a high-speed PCI Express 2.0 bus and cache memory has been doubled to 8GB of DDR3.

The controller units are available with 12 3.5in or 24 2.5in drive bays and you can pick and choose from 6Gbit/s near-line SAS, SAS and SSDs. Storage expansion has been doubled as nine disk expansion shelves are supported and up to 240 2.5in drives. For interfaces, you can choose from 6Gbit/s SAS, Gigabit or 10-Gigabit Ethernet IP SAN, 8Gbit/s FC and even 10GbE FCoE and mix different types of modules in the same RAID controller.

We were supplied with two separate head units each with dual RAID controllers all fitted with two 8Gbit/s FC interfaces. This setup allowed us to look at Fujitsu's new Remote Equivalent Copy (REC) feature which mirrors data between appliances in primary and secondary sites for disaster recovery. Each chassis was also each supplied with four 200GB Stec ZeusIOPS SSDs and twelve 300GB SAS2 SFF hard disks. These are included in the price we've shown making this system very good value.

The web interface has a quick start wizard for creating RAID groups. You can create multiple volumes within an array where each one is assigned a LUN and then mapped to selected interface ports. You can allow unrestricted host access to LUNs or use affinity groups which combine multiple volumes and only permit selected hosts to access them over specific ports.

Along with standard and thin provisioned volumes which are for general usage, you can create SDVs (snap data volumes). These are used by the SnapOPC+ (one point copy plus) feature which takes point-in-time copies of source volumes to an SDV. These copies are for temporary use, such as for offloading backups. You can map the SDV to another server where it initially appears as a raw disk. Once the SnapOPC+ task has finished, the SDV then becomes identical to the source and can be used for backup operations.

The REC feature mirrors data across separate arrays, but needs the additional SF Express software to configure and manage this.

QuickOPC can also handle long-term backups by making an initial volume copy and then updating it incrementally on subsequent copies. Equivalent Copy maintains mirrors of one volume to another on the same local array and Remote Equivalent Copy (REC) mirrors volumes across arrays in different physical locations.

REC is the one function that cannot be managed from an array's own web interface and this requires an extra Windows system running Fujitsu's SF Express software. This presents a separate web interface for managing multiple arrays. Once the controllers have all been registered with SF Express, you can monitor them from one location and create or modify RAID groups and volumes. REC sessions require two appliances to be paired together after which the volumes can be mirrored.

To test raw performance we used the Iometer utility configured with four disk workers, a 64KB transfer request size and ten outstanding I/Os. For the RAID-5 array we saw a superb raw read throughput of 783MB/s for one server and a cumulative 1565MB/s for both servers showing no contention for resources was occurring.

One can mix and match different types of interface modules in the same RAID controller.

We reran these tests on the SSD array and saw virtually identical speeds. However, using ten outstanding I/Os equates to light disk usage so we upped the pressure by changing this Iometer value to 256 to simulate very heavy usage.

Running the tests again on the RAID 5 array saw Iometer report a lower read throughput of 692MB/s for one server and a cumulative 1360MB/s for both servers. Moving over to the SSD array saw read speeds almost unchanged with one server returning 775MB/s which increased to 1560MB/s with both servers in the mix.

Along with excellent performance, the Eternus DX90 S2 is also very good value. Comparing the review system's specification with IBM's StorWize v7000 shows Fujitsu is offering a better package for a lot less cash and made all the more impressive as it includes eight 200GB Stec SSDs.

We found the DX90 S2 easy to deploy with high levels of redundancy and excellent expansion potential. Fujitsu will also be implementing automated data tiering towards the end of this year, so it will be able to dynamically move data between SSD and HDD storage based on usage patterns.

So what's our verdict?

Verdict

The Eternus DX90 S2 delivers on its promises as it is a highly expandable and, as our lab tests show, extremely fast storage array. The ability to mix different data interfaces on each controller makes it very versatile and the optional thin provisioning is a much needed feature. It's also good value as the price for our review system includes eight 200GB high-performance SSDs.

Two controller chassis each with the following specification: Chassis: 2U rack with 24 SFF hot-swap drive bays Storage: 4 x 200GB Stec ZeusIOPS SSDs, 12 x 300GB Seagate SAS 2 SSF hard disks RAID: 2 x hot-swap controllers with 8GB DDR3 cache Array support: RAID 0, 1, 10, 5, 6, 50, JBOD, hot-swap, global hot-spare Interface ports: 2 x 8Gbps FC ports per controller (max 4) Network: 1 x Gigabit Ethernet management port per controller Other: 1 x SAS expansion port per controller Power: 2 x hot-plug supplies Management: Web browser Software: Fujitsu SF Express

EMC may have just been dumped by Dell over a year earlier than expected, but its results show the storage giant is in fine health.

The company posted a 28 per cent rise in profits, following strong demand for its storage products and plenty of business from its virtualisation-focused subsidiary VMware.

Net income went up to $605.6 million up from $472.5 million in the same quarter last year, whilst revenue was up 18 per cent.

"The priorities we have outlined in our financial triple play' to gain market share, invest aggressively to take full advantage of the massive opportunities at the intersection of cloud computing and Big Data, and improve profitability continue to guide us through 2011," said David Goulden, EMC executive vice president and chief financial officer.

"We are confident that our company strategy, disciplined investment approach and continued focus on execution, which have driven our success over the past several years, will continue to serve us well in the future."

The break-up

The results came as EMC was dumped by long-time storage partner Dell ahead of schedule. The two were supposed to remain partners until 2013, but Dell decided this week it was not going to sell EMC storage products anymore.

Dell has its own storage products following the acquisition of Compellent, which was completed earlier this year.

Michael Dell's company was selling a wide range of EMC-made equipment over the past 10 years, including the popular CLARiiON and Celerra networked storage lineup, as well as NAS gateways and backup disk arrays.

The early parting should be good news for both parties. Dell is looking to become a complete data centre hardware provider, having bought a tonne of related companies on top of Compellent, including EqualLogic and more recently Force10 Networks.

EMC is looking to focus more heavily on its VNX products, moving away from its CLARiiON and Celerra products.

Just over a year ago, Michael Dell didn't have the full confidence of his company's shareholders.

A quarter of them withheld voting for the founder's re-election, as Mr Dell received less support than any other executive re-elected to the board.

The central issue facing the chairman and CEO was over a US Securities and Exchange Commission (SEC) investigation into how Mr Dell and his firm failed to adequately disclose information to shareholders over payments from Intel. The chip maker effectively paid Dell to keep using its processors, rather than switch to AMD.

The payments from Intel made it seem as though Dell was meeting or surpassing results expectations, the SEC said. Without them, however, the company would not have done so once between May 2001 and January 2006, the regulatory body claimed.

Dell was told to pay $100 million as part of the settlement, whilst the company founder personally paid out $4 million. It was not a pretty end to a murky story.

The legal investigation was not Dell's only problem. Some were calling for fresh blood, for a chief who could take Dell into a new era of IT in the cloud and thinner clients, as the desktop market falls.

This all came just after three years of Michael Dell's return as CEO of the company he founded, having stepped down in 2004.

Like a nerdy kid hitting the gym in earnest, Dell has bulked up to become a full data centre solution company.

He has made some bold attempts to answer questions surrounding his ability to take Dell forward over the last year, however. He has re-envisioned his company, whilst keeping a firm grip on the client market, as others namely HP have chosen to move away.

Has the company-coined "new Dell" brought him back from the brink and secured his place as the leader of the business he created?

Welcome back

From the outside, it appears so. Dell has been a confident company of late.

Back in February, it posted some solid results, as its fourth quarter smashed Wall Street expectations. Many pointed to a Dell revival, as the company shifted into higher margin business, including a greater focus on servers, storage and services.

In August, the company again posted some decent figures as net income went up 63 per cent year-over-year. However, it reduced its revenue expectations for the year down to between one and five per cent. Yet it would be harsh to slam Dell for this mixed showing, given the entire market is being hit by reduced demand, added to the fact it is going through a major transitional period. We can forgive this faltering for now.

Highlighting its commitment to change, Dell has spent a penny or two. As part of its redefining of its business to focus on overall solutions, moving away from a hardware-centric model, Dell has bought a raft of companies.

The acquisition of Perot Systems for around $3.9 billion kicked off the spending spree in 2009, marking a notable moment in Dell's services commitment. Then came KACE Systems in 2010, providing hardware-based systems management appliances, followed by storage company Compellent, having missed out on 3PAR. Its latest move was to get hold of Force10 Networks to provide some prowess in the networking space. Like a nerdy kid hitting the gym in earnest, Dell has bulked up to become a full data centre solution company.

"Finally, some of the power brokers in Dell have accepted that pile them high, sell them cheap' has had its day, and the acquisitions Dell has been making are all aimed at moving it into a much more complex, full hardware stack sell, with the software coming from solid relationships, mainly with Microsoft, but also on the Linux side as well," said Clive Longbottom, founder of analyst firm Quocirca.

"I do believe he has now seen that Dell could not go on as it was. New blood has been brought in, new approaches are being brought to the fore, and Dell is looking stronger than it has done for a while."

Lingering issues

On the face of it, Mr Dell has helped create a new company ready for today's IT world. Yet significant stumbling blocks remain. Number one is the declining PC market and Dell's falling position within it.

During its inaugural Dell World conference in Austin last week, Lenovo became the number two PC seller, pushing the US firm down into third place. Given the PC industry is already in decline, the Texan firm will not want to fall from the upper echelons of the market.

Exacerbating things is Dell's performance in other client segments. In tablets, one of the causes of the PC market's contraction, Dell has failed to impress. Its lineup of Streak devices have performed poorly, with the Streak 5 model ditched earlier this year. It has plans for a global release of 10-inch consumer-focused tablets, but they have not emerged in the West yet.

Dell hasn't made much of a mark on the smartphone space either. Despite producing a quality phone with the Dell Venue Pro, its market share remains infinitesimal. According to Gartner, its not even in the top 10.

Michael Dell may have to face up to the reality it shouldn't bet on the consumer segment as much as it currently is. Cisco has proven how disastrous major gambles in the consumer sphere can be, especially if you don't have the relevant acumen or experience.

By turning his ship towards a cloudy horizon and building hardware for the entire stack, Dell is looking like a considerably more attractive proposition.

In tablets, Dell will have to create something with top-level design quality and some significant power, as well as a low-entry price point if it wants to have any chance at succeeding. Whilst it can certainly create quality hardware, offering low costs looks unlikely. For starters, it won't be able to rely on apps and software sales like Amazon can.

Furthermore, no matter how much Dell says smartphones are part of the end-to-end solution for businesses, it can't deny the devices are primarily targeted at consumers. When you think of smartphones though, you do not think of Dell.

Simply looking at the most recent results, which showed much of the positives were coming from the enterprise segment, it should be clear to Michael Dell that placing considerable emphasis on the consumer is a big risk.

Let's get positive

Yet Mr Dell has done well. By turning his ship towards a cloudy horizon and building hardware for the entire stack, Dell is looking like a considerably more attractive proposition for companies who want a full business package.

"As long as Dell retains focus and manages to keep fighting with its broadened arsenal of compute, storage, management, network hardware, brought through with professional services and being sold as a solution sell, rather than a portfolio sell, I believe that it will begin to take business away from a wounded HP, and could also chip away at IBM where some organisations see IBM as being too expensive and cumbersome to deal with," Longbottom told IT Pro.

Michael Dell's second reign as CEO won't be of Steve Jobs proportions. He will not turn Dell into a cult-inspiring, behemoth of a tech provider, but there is little doubt he has steered the business in the right direction. All he has to do now is keep firmly on course hopefully steering away from consumer clients - amidst turbulent seas.

As IT spending looks set to diminish over the next year, Dell will have to look to those markets it knows will provide it with solid income.

Dell made a slew of mini-announcements around integration of recent acquisitions today.

Over the past year and a half, Dell has made some major purchases, including storage company Compellent, networking firm Force10 and data management provider Ocarina Networks.

All have formed part of Dell's move to provide the full end-to-end solution for customers, not just focus on hardware devices.

"Customers wanted us to know more about their businesses," said company CEO Michael Dell, during the company's inaugural Dell World event in Austin, Texas.

"You're going to hear about the new Dell which is an end to end service provider and yes we still care about the client device."

As part of this, Dell said it will add Compellent and Force10 capabilities to its vStart virtualisation offering, in which the vendor pre-packages and pre-tests a virtualised set-up before handing it over to businesses. These additions will arrive in 2012.

Dell also announced compression for its Dell DX Object Storage Platform using acquired technology from Ocarina, bought in 2010. The vendor promised the ability to compress stored data by up to 90 per cent.

This capability will arrive on 18 October, in the Dell DX6000G Storage Compression Node. It will be compatible with DX6000 environments, designed for storage of fixed digital content. The DX6000G will be an optional addition for DX6000 boxes.

The DX platform is designed to be "massively scalable" by turning files into objects, giving them metadata and then letting IT teams put policies on top of that. It is specifically designed for archiving of unstructured data.

Today, the company also confirmed a Windows 7 tablet would be coming this month.

ANALYSIS The mix of fixed problems and new features in Windows Server 8 is what we're used to in a new release from Microsoft.

There are all the features we expected for building private and hybrid cloud systems for enterprises. There's some significant improvements in virtualisation too that give it many more of the features of a server operating system.

Hyper-V can take advantage of a lot more hardware than before; not only can you use up to 160 processors - either cores or threads - and up to 2TB of RAM. Each virtual machine can have up to 32 virtual processors and 512GB of memory.

The new VHDX format supports files up to 16TB rather than 2TB and takes advantage of new sector sizes on the latest hard drives. There's no fixed ratio between logical and virtual processors anymore - you can run as many processors or workloads per physical core as the hardware can support.

Adding more hardware to virtualised systems doesn't always improve performance if the systems aren't using memory that's physically close to the processor thread. Rather than VMs just picking whatever memory is available, those that run on Windows Server 8 now get NUMA (non uniform memory access) support so performance will scale well as you add more cores to a server. It effectively means smarter memory management and therefore improved performance in theory of course.

Windows Server 8 also improves storage performance for virtualisation. If you have a storage array that supports offloading data transfer, when you move data inside a VM that can happen directly on the storage array. It means the system won't have to copy all the data across the network to the server and then back to the same array just move it.

It's easier to use virtualisation for making your infrastructure more reliable, like using Hyper-V Replica to take Volume Shadow Copy off site and cloning domain controllers so you can get them up and running quickly for failover and disaster recovery. Hyper-V also gets more operating system features, like whole-disk encryption with BitLocker even on clusters and using the Windows Hardware Error Architecture to look for problems like memory errors that cause page faults. Predictive failure analysis even finds some faults before they cause problems - because the memory is virtualised, the VM can just remap it without restarting.

Moving running virtual machines is also a lot easier. With Windows Server 8 you can move any part of a VM at any time, whether it's the whole VM, the configuration file for the VM, the storage that the VM is using or just a virtual hard drive inside the VM - and take the metadata and historical performance metrics with it. You can move a VM while it's copying files or running a computation (the work in progress gets mirrored until the migration is complete) and you can do all that over the local network, without stared storage or a cluster.

There have been rumours that Hyper-V was in maintenance mode' and Microsoft was concentrating on management tools and System Center. This array of significant new features for virtualisation proves that's far from true.

PowerShell powered

There are some significant improvements in management tools, although they're not always coherent or consistent unless you're ready to take advantage of PowerShell. In the messy real world the idealised clear and simple tasks of Server Manager aren't always what you need to do. PowerShell lets you get down and dirty but you still have high-level abstractions for tasks that work remotely. It's a powerful combination but even with the helpful new tools it still has a learning curve.

Microsoft is keen to get admins away from working in a GUI on a live server, encouraging them to use Server Manager or remote PowerShell from a client PC (or a Linux box). An interface is something else for the server CPU to run and something else that needs patching (and something you certainly won't get on a cloud service). There's even a new role between the existing command-line only Server Core, which is what Microsoft thinks you should run, and the full server role, or what Microsoft is calling Full Server without GUI.

Slightly confusingly you can still run Server Manager and MMC tools, but you don't get Explorer or Internet Explorer, making for a more secure system that's harder to mess up by accident. It's also easier to change the server role; if you need something that's not in Server Core you can switch to a Full Server role without installing the server from scratch, and you can switch back handy if you need the GUI to install an app but not to run and manage it.

But Windows Server 8 also looks forward to the next stage of commoditisation where x86 servers replace even more specialised storage and networking hardware. Storage Spaces let you make a pool of storage from SATA and SAS drives in a standard enclosure that you manage with standard Windows tools instead of needing to invest in expensive fibre channel storage arrays to be able to set up clustering.

Being able to manage thin provisioning with PowerShell and WMI opens up huge possibilities for virtualising storage on a budget. Clusters are much easier to work with - almost any server application can use cluster shared volumes and you can apply an automated rolling update across a cluster so you can get through Patch Tuesday without downtime. That means you don't have to be an enterprise to be able to have your server available all the time.

File deduplication is a high-end storage option that several companies have tried and failed to make more mainstream. Having it built in Windows Server 8 with tools that look both simple and powerful is a big advantage.

Using roles and automatically applied tags to protect files on file shares is a quick way to do a basic version of the deperimeterised data security you've needed complex and expensive DLP security systems to provide. You can create virtual network switches and run third-party extensions that let you monitor, filter and forward traffic so you can manage network traffic or change your core routing algorithms. These are features you tend to find only on high-end, high-priced network switches.

Essentially Windows Server 8 lets you take an entire network infrastructure, complete with security rules and provisioning policy and virtualise it on your servers which could save you a lot of money on network hardware and power without losing performance. Think about running a single converged network that handles storage, management, communications and server workloads.

The biggest concern we have is whether, even with the new HTML5 and WebSockets support, Internet Information Services (IIS) 8 has enough features to compete with open source web servers when it comes to powering the web apps that smartphones and tablets make increasingly more relevant. However, integration with existing server systems does make it a compelling solution for an enterprise that cares more about business web apps. Vastly improved options for managing security certificates may mean we'll see more web hosters offering IIS.

The underlying trend is towards putting more processing on servers, whether that's for thin clients, VDI or web apps. New options for making sure that server resources are shared in a fairer fashion whether that's between competing websites or between VDI users will make it easier to offer SLAs and do chargeback to make sure that departments get what they pay for - and pay for what they get.

A new H.264 codec, software GPU and support for touch and new USB devices (including USB 3) mean remote desktop sessions will get the full Windows user experience including the new Windows 8 features.

Whatever happens on the desktop and mobile side of the PC industry, Windows Server 8 is crammed with features that make it a strong successor to a server OS that already has 76 per cent of the market. Although it faces competition from software-as-a-service and cloud offerings, there are few businesses that will be able to manage without any servers at all.

While many of the features give mid-range companies access to options previously accessible only on large enterprise budgets, tools like the new, simpler DirectAccess (remote access without VPN) and peer-to-peer BranchCache are compelling for the smallest of offices.

As usual though, the biggest competition for Windows Server 8 is the huge installed base of existing Windows Server licences. It'll take time for any serious migration to happen.

Network storage arrays touted as being affordable for SMBs usually have limited expansion potential. In contrast, HP's new StorageWorks P2000 G2 arrays are designed so you can start small to keep costs down and grow easily with demand. They also have flexibility as their middle name as you can pick from controllers with 8Gbit/s Fibre Channel (FC), 1GB iSCSI, a combination of 8Gbps FC and iSCSI, 6Gbps SAS or the latest model on review here with dual 10-Gigabit iSCSI.

HP's 2U chassis supports either 12 3.5in (LFF) or 24 2.5in (SFF) hard disks and you can choose from 6Gbit/ SAS, midline SAS or low-cost, high capacity SATA. For the latter HP advised us it expects to have 3TB drives certified for use by late autumn.

The chassis accept a pair of hot-plug controllers so you just need to pick which interfaces you want. You can also reduce costs further by starting with one controller and adding a second when you need it.

The controllers have 2GB of cache memory and use a capacitor and flash memory to protect the cache contents. If there's a power failure, the cache contents are written to flash memory and the main advantage here is the capacitor takes far less time to charge than a battery.

The controllers have 2GB of cache memory and use a capacitor and flash memory to protect the cache contents.

For installation you can start by accessing the command line interface (CLI) using the mini USB port on the controller and the supplied USB cable. This will require a proprietary driver which can be downloaded from HP's support site, but note that it does not support 64-bit Windows systems.

We found it easier to point a web browser at the controller's default IP address where we were greeted by a nicely designed interface. It opens with graphics of the array chassis, virtual disks and all controllers and you can drill down to view individual components.

Expansion units are available with 12 LFF or 24 SFF bays and use simple controllers with 6Gbit/s SAS interconnects. This is where the P2000 G3 really shines as the controllers support a maximum of 96 LFF drives or 149 SFF drives so expansion potential is extremely good.

Your first job is to create virtual disks, or vdisks, where you select physical drives, pick a RAID array and assign hot-spares. Make sure you choose an appropriate array type as you can't migrate from one to another although you can expand existing arrays by adding more drives to them.

A provisioning wizard helps with vdisk creation and can also spilt a vdisk into equally sized volumes if you want.

Vdisk creation can be speeded up with automatic setup but with full manual control you can choose an array, assign a primary controller to it and decide on the number of volumes it should contain. You can leave the controller to create equal sized volumes where you simply name them and dish out host mappings to define host access privileges.

You can use snapshots to create point-in-time copies of selected volumes which are stored in a separate pool on the appliance. The base price includes a 64 snapshot license which can be upgraded to a maximum of 512, a very useful extra that provides a lot of added value.

There's more as you also get volume cloning as standard. This is used to create an identical copy of a volume which can then be used for a variety of purposes such as archiving data or safely testing applications. It's as easy to use as snapshots - simply select a volume, choose the copy option from the drop-down menu and either run it immediately or schedule it for later.

HP supplied our review unit with eight 1TB SATA drives which we used to create a single RAID-5 array. The build process took a few hours to complete, but providing you select the online initialisation option you can access the array during this phase.

For performance testing we used a Broadberry dual 2.8GHz X5560 Xeon server running Windows Server 2008 R2 64-bit and equipped with an Intel 10GBase-SR card. We used this to directly connect the server to the primary controller and had no problem logging on using the Microsoft iSCSI initiator where we could see all mapped volumes.

Once the RAID-5 array had initialised we mapped a volume to the server and used the Iometer utility to test raw throughput. Using eight disk workers and 256KB transfer request sizes we saw an impressive top speed for sequential read operations of 676MB/s. Changing Iometer to 100 per cent sequential write operations returned an average speed of 465MB/s.

A Mini USB port on the rear of the P2000 allows you to control the storage array using a command line interface if you wish.

Along with fine performance, the new 10 Gigabit Ethernet model of the P2000 G3 comes in at a price that few other storage vendors will be able to match. We found it extremely easy to use making storage provision a simple task and the high expansion potential makes it a good long term investment.

So what's our verdict?

Verdict

The StorageWorks P2000 G3 is a great choice for SMBs looking for a flexible network storage solution that won’t break the bank. HP offers an excellent range of controller choices, expansion potential is very high and the price includes snapshots and volume cloning. The smart management interface makes light work of storage provisioning and the new 10GbE model on review performed very well in our tests.

Chassis: 2U rack Storage: 12 x 3.5in drive bays supporting SAS, midline SAS and SATA hard disks Power: 2 x 573W hot-plug supplies Two controllers each with the following: Memory: 2GB cache with capacitor and flash memory backup RAID support: RAID 0, 1, 3, 5, 6, 10, 50 Data ports: 2 x 10 Gigabit SFPs Management: Fast Ethernet port Other ports: Mini USB CLI port Management: Web browser Snapshots: 64/512 (Standard/Maximum) Volume Copy: Included as standard Options: AP843A P2000 LFF expansion enclosure, £2,259; AJ941A D2700 SFF expansion enclosure, £2,319 (all ex VAT) Model code: AW596A

Storage company Drobo has announced a new range of redundant storage arrays aimed at businesses, along with the after-sales care to back them up.

Drobo's eponymous multi-disk storage arrays used a RAID-like technology, branded BeyondRAID, to ensure data redundancy but, unlike traditional RAID arrays, can consist of disks of any size and are easier to set up and administer. Many previous Drobo arrays have been aimed at consumers, but the new B series models are aimed specifically at businesses.

The massive B1200i is the most interesting since it has iSCSI support and can accommodate up to 12 disks. It supports tiered storage, so frequently accessed data is stored on the fastest available disk, as well as thin provisioning. It's rack-mountable and has dual, hot-swappable power supplies.

The smaller Drobo B800i and B800fs can take up to eight disks, but the B800fs drops iSCSI support. The B800 models will accept SATA hard disks in any combination of capacities. Oddly, Drobo claims that only the B1200i can accommodate SSDs and SAS hard disks. All three models are backed up by the 24/7 DroboCare service which aims to deliver replacement parts on the same business day any fault is reported.

All three models can now be managed from the new Drobo Management console which aims to provide an animated, but logically organised administration interface whether your Drobos are located on another floor or in another country.

The B800-series Drobos should be available shortly with prices from 1,409, but the B1200i won't be available until April. We'll bring you full reviews as soon as we can.

If there's one thing businesses can be sure of it's that demand for network storage will never stop growing. However, despite being faced with this voracious appetite, administrators simply can't keep on feeding it with more hardware. They must instead manage existing resources more effectively to reduce wastage.

SRM (storage resource management) was always supposed to be the answer, but we remain unconvinced by many of these products as they can be expensive and overly complex with few capable of managing mixed vendor environments. Storage Profiler from SolarWinds aims to avoid all these problems as it's designed to be easy to use, good value and capable of working with a wide range of network storage hardware.

Network resources supported include file and backup servers, RAID arrays, HBAs, NAS appliances, IP and FC SANs and VMware environments. However, pricing can be complex as Profiler has a number of optional components with the base product capable of monitoring SANs, NAS appliances and HBAs from vendors such as NetApp, EMC, Cisco, HP/3PAR, Dell/EqualLogic and IBM.

Licensing is based on the number of physical hard disks to be monitored, so even if your NetApp array has one virtual drive, a license is required for each member disk. Profiler can also monitor FC switches from Cisco, Brocade and QLogic, but this costs an extra 690 per switch.

For Windows Servers and VMware systems you need the Virtual & Server Profiler. It includes agents for Windows systems which provide enhanced reporting on system utilisation and resident files. Profiler doesn't use SNMP to gather information about system utilisation, so if you want to know about CPU, memory, disk and network usage on your Windows servers you will need this agent. It instead uses the VMware API to query host systems about system utilisation, VM activity and storage usage. It also provides details on end-to-end mappings and the relationships between VMs, datastores, LUNs and physical arrays.

For testing we had all options licensed and found installation to be a simple task with the Profiler server loaded on a Windows Server 2008 system in a few minutes. It doesn't offer automated agent deployment so we installed these manually on our Windows Server 2003, Server 2008 and Server 2008 R2 test systems.

The console's home page opens with a basic summary of managed systems and provides plenty of links to help and tutorials. Profiler uses policies and rules to control most functions. One rule allows agents to auto-register with the server which we found occurred within seconds of the agent being loaded.

A network discovery tool looks for EMC, LSI, NetApp and VMware systems and a new feature is support for SMI-S. Our test environment also included a VMware ESX Server 4 system which Profiler found and, once we had provided login credentials, added it to the list of managed hosts. SNMP must be configured for VMware and we found SolarWinds provides excellent instructions for this.

The main console page has an overview of monitored resources with pie charts showing total usable and raw storage plus VMware datastore usage. Beneath this are tables showing the top ten servers with the least amount of storage capacity available along with the highest processor and memory usage.

Select any of these graphs and you can drill down for more information and eventually arrive at individual systems. This window shows disk, processor, memory and network performance graphs as well as a capacity analysis which lets you see if free space on each volume is declining or increasing.

The Backup Profiler can be difficult to configure as it requires a lot of information about the backup application including log file locations and data and time formats. We also found out the hard way that it doesn't currently support CA ARCserve r15. We had more luck with Backup Exec r12.5 Profiler successfully detected two servers running this software. The Backup Monitor window kept track of successful, running and failed backup jobs on these backup servers.

Add servers to Backup Groups and you can view graphs of total backed up data, throughput and the percentage of successful jobs. If a backup job fails, the icon for the relevant server turns red to highlight this immediately. The event monitor keeps you updated on any problems with colour coded alerts and, once profiles have been configured, you can request alerts on issues such as capacity shortages, high processor usage and so on.

Profiler comes with hundreds of predefined reports covering all areas from volume usage for systems and groups to backup operations or available storage assets to SAN topologies. You can also use any of these as templates, customise them to suit and have them emailed out at regular intervals.

Even though some of the optional modules will push up the price, we still think that SolarWinds Storage Profiler is comparatively good value. It supports an extensive range of hardware vendors, is relatively easy to deploy and is capable of providing a wealth of information about your storage network.

So what's our verdict?

Verdict

If you’re not sure what’s on your storage network then we recommend checking out Storage Profiler as it’s capable of providing a lot of useful information about an extensive range of devices. You will need the optional modules to get the best out of it, but this software will make sure you never run out of storage space again.

SYSTEM REQUIREMENTS Memory: 2GB Hard disk: 50GB Operating system: Windows Server 2003, Server 2008, Server 2008 R2 (32 and 64-bit) Options: Virtual & Server Profiler – £1,315 for up to 10 virtual or physical servers, £2,635 for up to 25 virtual or physical servers. Backup Profiler per primary backup server – Symantec Backup Exec, £350; CA ARCserve - £520, Symantec NetBackup - £3,450.