On Premise

Good news and bad news today: the good news is that people are actually reading this blog! Or at least, that’s what they tell me. The bad news is that things are really busy in the office… so blogging has to happen in between other tasks. But anyway, here we go: post number 4 (four).

One of my many readers (j/k) asked me my view on EMC’s “On Premise Cloud Solution”. Being here to serve the community – and probably also a bit because I’m curious myself, I dug a bit deeper into the world of “On Premise Cloud Computing”. Mind that this is a marketing blog – not a Geek Speak blog so expect no technical deep dive!

The past few months I had the impression that most people agree on three levels of Cloud Suppliers:

• The Infrastructure guys (guys with data centers)
• The Platform guys (guys who put a cloud platform on a data center and give APIs to the Apps guys so that they can do their stuff)
• The Apps guys (guys who create so-called cloud applications)

* For more details on those cloud levels, read this and this.

As the Apps guys are closest to everything that is Web 2.0 and easiest to explain to the large audience – but also because some believe that this is where money will be made – they are really getting the bulk of the attention these days. And so do the Platform guys as they enable the Apps guys. But what happens on the infrastructure level? Who is building clouds? Or is there no more need to build more clouds? Of course there is: we need more public clouds AND there is a big demand for private clouds. According to this blog, 2009 will be the year of the truth [quote]: “Traditional enterprise will want to have “hybrid clouds”; virtual infrastructure in their private data centers connected to cloud facilities for “fill in the blank” reasons”.

So, apart from Q-layer – who obviously have the best solution to build private clouds (J/K) – who else can build On Premise clouds? EMC claim they do… Let’s have a look:

Ironically the first site that informs me on EMC’s offering is Azurejournal, who are relatively positive on Atmos, EMC’s cloud solution. But they also call it an “information management solution”, which made me a bit suspicious. And yes, on the EMC website, it becomes all very clear: I was wrong – it is a cloud solution, but it does not build a cloud in the sense of “one big pool of data center resources.” It is a cloud storage solution: “an information management solution designed to help customers automatically manage and optimize the distribution of information across global, cloud storage environments”. And as EMC is pretty good at storage solutions, it’s probably not a bad product.

And Atmos does have a couple of very nice features. For example, the policies that define how many copies of files need to be kept available depending on the popularity of those files. Of course they also built-in deduplication and a range of time saving administration tools. Cherry of the pie (naturally) is a rich set of API’s to enable commercial models. No surprise that Atmos supports VMware but what about Hyper-V? Isn’t Microsoft still very powerful in Corporate and Governments? What’s also nice is that EMC claims this is a 100% in-house development and it has been used by customers for half a year. Unfortunately, the customers wish to remain undisclosed. What “sounded” as the nicest thing about Atmos is that the product is hardware agnostic, which is courageous for a storage vendor, but they immediately add that there are “recommendations”. Nice try 😉

Based on the marketing collateral out there, I think it is safe to say that EMC made a nice product, but I’m expecting a lot more On Premise Cloud solutions to be brought to the market and only then we will be able to properly evaluate this product. So as a conclusion, I think that in spite of all the attention going to cloud application providers for public clouds, there is still a lot to be expected for the enterprise and government markets. There seems to be a consensus that these markets will act in two phases: first build private clouds and then connect those clouds to public clouds and thus create what we call hybrid clouds. But as Chuck Hollis says, it is early and “For those of you consuming traditional storage with traditional IT use cases, none of this will matter much to you in the short term.” After all, it’s probably way too expensive for innovators anyway.

I can’t way to see this happen (in 09)!

Oh and for those guys who really want the technical deep-dive, I really enjoyed this.

~ by tomleyden on December 11, 2008.