Along with support for low-cost SATA SSDs, it claims a high throughput of up to 600K IOPS for 4K random writes with low latency while its innovative 'share-nothing' architecture and high availability (HA) clustering services have no single point of failure and deliver an industry-leading six 9s reliability.

Value looks even better as AccelStor systems come out of the box with everything enabled so there are no hidden costs. Features are extensive and include thin-provisioning, deduplication, compression, fast point-in-time snapshots, cloning, support for SEDs (self-encrypting drives) and remote replication for off-site disaster recovery.

A unique feature of the H710 is AccelStor's patented FlexiRemap technology which is designed to handle the heavy random I/O workloads that traditional RAID arrays are inherently unsuited to. Whereas RAID algorithms randomly write data to SSDs in an array, FlexiRemap reorganises them as sequential writes allowing it to evenly distribute them across all array members.

I/O intensive applications will see many benefits as FlexiRemap improves random write performance while minimising SSD wear. AccelStor claims FlexiRemap can increase storage array longevity by years and provides advanced data protection when compared to standard RAID arrays.

The H710 is delivered as well-specified 2U rack servers comprising two independent nodes supporting 8/16Gbps FC or 10GbE iSCSI operations. They run in symmetric active/active mode and are linked together over high-bandwidth Infiniband ports for storage pool synchronisation between the nodes.

Another key differentiator is that unlike many competing products that employ ALUA (asymmetric logical unit access) mode, the H710 functions in full symmetric mode. For the former, hosts that access volumes over a port on each controller will see one as active and the other reserved for failover whereas the ports on both H710 nodes are all active and can be aggregated into high-speed, fault-tolerant MPIO links.

Each H710 node has 24 SFF drive bays and our review systems were supplied with a full house of 1TB SATA SSDs. In HA mode, the nodes automatically function as a cluster where a single storage pool is mirrored between them - similar to a RAID 50 array.

Its well-designed web interface made the H710 easy to deploy and we had HA configured in a few minutes. We could then create multiple virtual volumes in the synchronised storage pool, assign LUNs, choose which data ports to present them on and configure manual and scheduled snapshots.

Snapshot backups are equally simple to configure and after creating a snapshot schedule on a virtual volume, we declared a CIFS share on a NAS appliance as a destination (NFS is also supported). You have a number of options as backups can be set to run the moment a snapshot has completed or they can be scheduled to run at a specific time each day. Snapshot backups to the cloud are also supported with options provided for sending them to Amazon Web Services (AWS) buckets.

Deduplication and compression are enabled on selected virtual volumes and set to run inline in the background while enhanced services offering higher space savings can be run to a schedule at off-peak times to minimise any performance impact. These features are very efficient as when enabled on one of our test volumes, they returned a high space reduction of nearly 60 percent.

Product: NeoSapphire H710
Supplier: AccelStor Ltd
Web site: www.accelstor.com
Sales: sales@accelstor.com