It was a two-beer argument. You know the sort – it’s the type of discussion that requires at least two beers before you can put away the trivial worries of raising a family, running a business or making the mortgage payment in favour of matters of great and unequalled importance. On this occasion, the two-beer argument triggered a bitter war of words around a table in a Soho pub with a few colleagues over the true meaning of the increasingly over-used phrase ‘software-defined storage’ (SDS). With a glut of vendors (both new and old) releasing software-defined solutions on a near-daily basis, it is a debate I am seeing take place on a regular basis at trade shows, meeting rooms, conference calls and, of course, the pub.
The most common definition generally outlines SDS as any solution that favours the use of software, rather than hardware, for the control and management of storage and data. On this point, we were all agreed. However, we could not find common ground on the finer points of what this actually means. Can there be an ASIC anywhere in the stack? Is there any hardware- based RAID involved? Furthermore, is there any RAID involved at all, hardware, software or otherwise? What if the software is only supported on one type of hardware?
We went in circles with this sort of drivel and no one participant could gain any discernible ground over the other, always focusing on the technical elements. Each new rebuttal was delivered just a little bit geekier than the last. But were we getting too bogged down in the details, ignoring the actual point of SDS in the first place?
One of the major and most important capabilities of SDS is that it brings you flexibility and agility. It can empower a business to expand, shrink or change their storage infrastructure on the y to keep up with rapidly evolving requirements. It can allow the purchaser to avoid lock-in by any one particular hardware vendor. It can provide the opportunity to make cost-conscious choices without sacrificing functionality or stability. And it can potentially do all of these things without creating unmanageable technology ‘islands’ and access points. Simply put, SDS can save you an enormous amount of money. Of course, this in itself does not imply a fixed definition of SDS, it is simply one of the achievable benefits that could potentially be realised from investing in a software-defined solution.
‘One of the major and most important capabilities of SDS is that it brings you flexibility and agility’
Therein lies the problem with the current hype surrounding SDS. It’s not the definition of SDS that matters, it’s the application of it. Listing SDS as a feature of a particular product or solution when you can only run the software in production on a prescribed hardware platform does not go very far in showcasing SDS as an actual benefit of the given offering. If your choices are limited, who cares if it’s software- defined or not?
In other words, being software-defined for the sake of being software-defined is nothing more than a feeble attempt at capitalising on the rapidly growing uptake of software-defined solutions as a whole. With all of this in mind, there are three questions you should ask any sales rep preaching about the software-defined elements of their offering. First, can I run your software on the hardware of my choice? Second, will you actually support me if I do? Finally (and this is the really important one), can I actually do this today? If the answer to any of the above is ‘no’, advise them to swiftly move on to the next section of the pitch.
At Pixit Media, we are absolutely fanatical about SDS. The foundations of our entire business model are firmly rooted in our ability to enable a huge amount of flexibility and choice for our customers and partners when making hardware purchasing decisions; we are able to do this because PixStor is software-defined. Our license costs are predictable, and because no single hardware vendor can ever claim to ‘own’ a PixStor installation, so too are the hardware costs when scaling out both the capacity and performance of PixStor. Of course, this does not make us any more or less software-defined than our competitors at a technical level. It simply means we are actually leveraging SDS to deliver real value both in the data centre and at the bank.
Back at the pub, we all eventually realised that we had taken this semi-lubricated game of verbal tennis totally in the wrong direction. The fact is, spending time arguing the technical aspects of software-defined ‘anything’ when attempting to crown one solution or the other truly software-defined is about as useful as pouring boiling water into a chocolate teapot. Without tangible business bene t, the SDS component of any solution is as irrelevant as the definition itself. So we stopped bickering, emptied our glasses, shook hands and went home to worry about our kids again.