Permabit just made a move that has the potential to be very interesting. It has taken its Alberio deduplication and compression abilities and packaged them into an appliance so as to be able to - essentially - retrofit data reduction capabilities to installed FC SANs. Of course data reduction is nothing new per se; various efforts have been made by the “mainstream” vendors for their “mainline” products over the last few years but without a huge success….although maybe because such efforts have not been wholeheartedly embraced at the sales tip of the spear, given that they almost certainly lead to lower capacity sales. However, as the saying goes, the times they are a’changing: Data reduction is cool, embraced and promoted by the newer vendors, and – with the cat thus firmly out of the bag – there is an undercurrent of pressure for it to be more widely available. There are few if any realistic reasons to not use it….well, except, ahem, for the fact that it ain’t available on most of the common products in use (and indeed still being sold) today.
Thus the new SANblox move by Permabit is intriguing - not to mention potentially lucrative - for a number of reasons:
- Dedupe has for the most part been, until now, a key ingredient in the special sauce by which the all-flash vendors (especially) get to say that they can get their product cost down to a level similar to spinning drives; if existing FC SANs can now easily, and at low risk, add that same function then the price delta could be expanded again.
- There’s a likely performance boost as much as the $$ motivation. And, also, of course the function will still work as and when users upgrade to other newer devices from their favorite vendors that have (oh yeah….) the Permabit software included.
- Permabit is a proven piece of software- also it has made installation really easy while also providing HA via synchronous writes to ensure data safety. Its own testing shows data reductions typically run in the 4-6X range.....in other words, for many workloads you might only need 15-25% of the storage space you thought you needed. That’s no small improvement when you look at the cost of storage systems!
- It has the ability to be a "pull" technology....as users get to know it can be done they might well exert pressure on their vendors to support it. Key products from major vendors such as EMC, NetApp, Dell, and Hitachi have already been qualified…..one cannot imagine that such traditional vendors are all 100% thrilled at the prospect of such pixie dust being sprinkled on their systems, but –equally – their pragmatism and desire for account retention could conceivably actually drive them to desire to sell less capacity!!
- Why would vendors do that? Well....
- they need more efficiency tools to stem and manage general storage growth; indeed getting more back-end efficiency might not translate to less revenue as users are likely to continue to spend the same budgets but be able to do more for those budgets. As always, there’s plenty of actual and nascent capacity growth to go around.
- it's a bit like things such as vVols from VMware. As a “traditional” vendor you might not like it but you have to be seen to be a part of the contemporary world.
- increasingly vendors are making - and going to make – more of their money from software and so squeezing more capacity out of the back end HDDs isn't as painful for them as it once might have been.
This is going to be fun to watch – the question marks over this are that some users simply don’t like dealing with smaller companies (but this will be an OEM sale so that mitigates that), and that the extent of uptake by the existing big FC array vendors will largely be driven by “have to” rather than “want to.” But our industry has an uncanny history of gravitating towards attractive financial propositions once they are available. And vendors have to tread carefully to maximize their value-proposition in that transition, balancing revenue (from selling existing capacities) with risk (from not offering the reduced capacity and annoying their customers). In this case it is “caveat vendor” just as much as the more usual caveat emptor!