Now, to be clear, Isilon does allow multiple separate workloads with separate IO, even separate storage pools. However, some questions arise about those pools. They’re not encrypted from one another with keys controlled by the customer, nor is it even clear that they support multitenancy. There are no integrated billing capabilities, and there is no integrated service level agreement (SLA) management. In other words, it does not appear to be truly cloud ready.
In my opinion, true cloud storage is disruptive, and the traditional monolithic big-box vendors are trying to obfuscate what a cloud really is, in the hope that people stop recognizing that the cloud is a disruptive paradigm shift.
As I mentioned in a previous blog, neither Amazon, Google, Nirvanix, Microsoft nor others are building their infrastructures on expensive hardware from the big monolithic storage box pushers. Yet they’re all somehow managing to attract plenty of new customers, both large and small in scale.
One theory I have is that some of these storage boxes with the “cloud” term appended to them are lousy at actual cloud services intentionally, as a way to put a sour taste in people’s mouths when it comes to cloud. These box vendors want the cloud to go away, because if the cloud really succeeds, it decimates their margins. Is it incompetence? Or is it a Machiavellian strategy? Are the monoliths really just sinister architects planning
for the downfall of the cloud? Only time will tell.
Fortunately, more and more companies are looking at storage like consuming electricity. They’re realizing that if they bought their electricity in the same way they bought their storage, life would be miserable. For example, instead of paying a monthly service bill to the electric company, they would have to buy a generator, keep buying fuel so it never runs out, pay for maintenance on the generator and buy a new generator when the old one wears out. Now, unless you’re a manufacturing plant that uses a ton of electromechanical machines, this scenario is economically unfeasible. And the same can be said about the current, aging model of acquiring storage.
