Taneja Group Analysis
Once upon a time, when clouds were young, we thought that all our fancy storage arrays would shortly sport built-in cloud tiering mechanisms. It seemed natural to want to burst capacity to the cloud, maybe even replicate some data to a cloud DR clone, or archive corporate data into some big data lake in the sky. There were a few gateway products and 3rd party replication utilities rolled out here and there, but for the most part on-premise storage and cloud storage have been kept a world apart. Sure, there were some security and performance bumps to smooth out in order to reliably hybridize important data, but what we think the last real problem was that SAN vendors just didn’t really want to flow capacity off-site (and off their books).
But now that trusted IT vendors like IBM have their own successful public clouds going strong, that final roadblock has come down. In fact, IBM Spectrum Virtualize, which is built into many IBM storage solutions (like Storwize and FlashSystem v9000) and available as a virtualizing solution over most 3rd party storage, is now inherently cloud-enabled. It can now natively migrate and/or replicate to the cloud, passing data back and forth between on-premise storage in its purview and IBM Spectrum Virtualize for Public Cloud running in the 25 world-wide IBM Cloud data centers. And to meet new multi-cloud approaches, it all works cloud data center to cloud data center as well.
Microsoft had explored a smaller version of tiering to cloud storage with their StorSimple array, but that’s now only used a gateway “drug” driving folks into Azure and Azure storage. IBM doesn’t have this overt “it’s all cloud” agenda, rather Spectrum Virtualize really opens up and unlocks hybrid architectures of all kinds, enables more cost effective archiving, big data scenarios, cloud-based DevOps, and perhaps most immediately, actually effective cloud-backed DR strategies for existing enterprise storage of all kinds (whatever can be virtualized by Spectrum Virtualize).
At this time IBM does steer you into using IBM cloud footprint, but to extend the life and capacity of relatively expensive on-premise storage assets, and actually protect enterprise operational block data off all (most?) arrays, we don’t think that’s much of a problem. If you have Spectrum Virtualize assets today in your data center, and aren’t otherwise taking advantage of cloud migration or replication, it’s now time. Go cloud!