IBM pumps up the volumes for data pool party

IBM has a new release of its SVC software coming out to widen the distance between cluster nodes, improve HA and provide better Tivoli integration. This strengthens the V7000 too, as it shares the same SVC binary code.

SVC is IBM's SAN Volume Controller, it is a block-accessed, network storage product that sits inline in a SAN fabric and virtualises IBM and third-party storage arrays behind it, presenting them as a single pool of storage. Of course it does more than that, with tiering of data across different tiers of storage, thin provisioning, clustered nodes, and so forth. The V7000 storage array uses two SVC controllers running exactly the same binary code as the SVC product itself.

No other supplier has anything like SVC, although HDS's VSP and USP-V have SVC-like functionality in their controllers. NetApp's V-Series can virtualise third-party arrays and present them as FAS unified storage arrays.

EMC had Invista, a back-end array virtualising blade that slotted onto SAN Fabric directors but was too limited in its functionality by the limited space it could occupy and never became successful. The company has latterly introduced VPLEX, a hardware-software product to virtualise and federate VMAX and other arrays, including non-EMC arrays. It has even been used to present a common pool of storage with no EMC storage behind it, just third party flash and disk drive arrays.

V6.3.0 improves SVC's high availability (HA), making it more reliable,and integrates SVC more closely with Tivoli Storage Productivity Centre.

Stretched clusters

The capability being added here is to have an SVC cluster with nodes many kilometres apart. IBM says it enables "enterprises to access and share a consistent view of data simultaneously across data centres, and to relocate data across disk array vendors and tiers, both inside and between data centres at full metro distances".

There is an obvious applicability to virtualised servers and data centres and the movement of virtual machines between servers within and between data centres.

Higher availability

SVC Global Mirror can use FlashCopy targets as sources in remote copy relationship. These are space-efficient and enable lower bandwidth remote mirroring than was previously possible. Such mirroring can be done between an SVC and a V7000. IBM states that remote deployments for disaster recovery for current SVC environments can easily be fitted with the V7000 or vice versa.

Another SVC improvement is the use of round robin scheduling to better manage the I/O paths between SVC nodes and attached servers. Native LDAP support has been added as has support for more third-party arrays such as the thrd-gen XIV, HP 3PAR, Bull StoreWay, Fujitsu Eternus, TMS RamSan and Violin Memory.

These software enhancements strengthen the SVC in its two roles of an in-fabric virtualiser of IBM and third-party storage arrays connected to the fabric, and as a virtualiser with its own storage as the V7000. With stretched clusters, IBM can use the SVC as its answer to VPLEX for metro storage node clustering. ®