IBM launches Spectrum Accelerate, corrals SDS under one name and organisation

I’ve been telling everyone for a long time that IBM’s the leader in Software Defined Storage, particularly for the hard work it’s been doing for so long to address other vendors’ arrays in its SAN Volume Controller (SVC). EMC may be the market leader in storage hardware now (see Figure), but IBM has 700 storage related patents and invented tape and disk drives. It announced its ‘Spectrum’ SDS brand, launched the new Spectrum Acclerate offering and a number of statements of direction in the storage area this week.
You’ll want to know more about what it all means.

I think this is an important announcement, which fixes 3 related issues. In particular:

• Its storage developments were disparate, despite various movements of staff from Software to Systems groups in recent years; it makes sense to have a single ‘Storage and Software Defined Systems’ team headed up by Jamie Thomas and a single sales force under Richard Smith; its proposed $1b spending over the next 5 years will be better used than before as a result
• It’s naming schema was awful; O.K. if you’re storage experts speaking in a secret language you don’t want anyone else to understand, but horrendously difficult for business managers and CXOs to understand; pulling these products together under the ‘Spectrum’ banner will help IBM and its customers get focused on SDS
• In the past IBM – like most large storage systems suppliers – was very guarded in setting its Intellectual Property free from its arrays; Its latest announcement underlines a new intention to sell software-only products alongside versions tied into arrays as well as in ‘as a Service’ Cloud offerings

IBM Spectrum Accelerate is new and allows a number of intelligent features derived from XIV to be used in (initially only) VMware environments. These include its ‘zero-tuning’ architecture for adding storage capacity rapidly – a massive time saving over installing and implementing traditional arrays. It also provides business continuity for all committed data on disaster – better than losing 15 minutes or so of data with some of its competitors’ software offerings. IBM is using a TB pricing model for Spectrum Accelerate and intends to launch it ‘as a Service’ on SoftLayer sometime in the future. XIV is an object-level storage array used most famously by Netflix. Although not available at launch IBM will enhance Spectrum Accelerate to be able to handle VMware Virtual Volumes later.

Table: IBM Spectrum Software Defined Storage products at a glance

Source: ITCandor (after IBM)

I’ve listed the other Spectrum offerings in the Table; each one is related to – and sometimes no more than a renaming of – its earlier offerings. This is the state of play with each:

• Spectrum Archive is ‘network attached unstructured data storage’ supporting LTFS, which works in conjunction with Spectrum Scale; it creates an active archive, with tape as a tier of storage as opposed to an offline repository
• Spectrum Control is a virtualization platform and management solution for Cloud-based and Software Defined Storage; it has built-in Cognos analytics to help with the placement and movement of data; it can be used for managing heterogeneous environments; it is available as a vCenter plug-in; it also announced a beta version of Spectrum Control Storage Insights, which is a Cloud service running on SoftLayer
• Spectrum Protect is a data protection and recovery solution for physical, virtual and Cloud data; it has de-duplication built in
• Spectrum Scale is a data management solution for scale-out file and object storage; it has creates a single global namespace allowing a single file system to run across multiple sites; it is POSIX compliant, and so is integrated with OpenStack (for which it has Cinder and Swift object storage drivers) and Hadoop; enhanced encryption was added last year
• Spectrum Virtualize is a scalable storage virtualization system I prefer to call a ‘storage hypervisor’; it pools arrays from IBM and most competitors, has compression built in and eases the movement and management of data typically in large organisations

IBM also announced its intention to introduce its Multi-Cloud Storage Gateway later in 2015. This will allow Cloud storage to be managed as a storage tier, tying together on premise and on Cloud with advanced functions such as compression and encryption. Users will be able to fan out highly secure data across multiple Cloud instances running on Rackspace, Amazon S3 and/or Microsoft Azure.

While keeping and protecting organisational data is of prime importance, storage admins are no longer the most conservative of the IT staff and all major system vendors – IBM included – were caught short by changing demands. Users now need to address speed and capacity issues differently at times, use commodity hardware where appropriate and pool all of their resources as much as possible. For many ‘Software Defined Storage’ is a rallying cry: it means unifying the hardware, federating usage with a variety of storage vendors’ equipment, integrating with leading virtualized environments and addressing hybrid on, off premise and Cloud storage environments.

IBM’s been listening and has made some big changes in the way it runs its storage as a result. This is a massively strategic announcement for IBM and signals its intention to turn its storage expertise inside out from being a feature tied to its hardware arrays to software-only licensing and the Cloud.

Further reading:

IBM Storage

IBM and VMware