Next Generation Data Center – 4 years forward 3 years behind

0
293 views

Share on LinkedIn

I attended an Oracle “Next Generation Data centre Summit”, because I was interested to see how Oracle presented their SUN acquisition, and to see how they positioned Cloud given that Larry Ellison is often represented as being a Cloud sceptic.

Oracle now presents itself as from “Disc to Application“, meaning they are an in-house shop from SUN’s disc drives upwards to the Oracle business applications. The way they have driven their core database products to higher performance through very tight integration with SUN’s hardware, OS and middleware expertise IS very impressive. Oracle is a powerful beast!



I’m not so sure that being a totally integrated company is inherently a good thing. After all that’s what all the giant IT firms were like 30 years ago, and with the exception of HP who degenerated back into it the others have all divested.

Back to the theme – the Next Generation data centre – the core message was that it takes a lot of planning and a lot of effort and skill to re architect a new data centre. Oracle (Sun) sees it as the progression – from a physical centre, to a virtualised centre, an “Internet” centre, and then a Cloud data centre. They say most people are at the virtualised or Internet stage, and it’s about stepping to the Cloud generation.

So this reveals Oracle’s “Cloud” – selling companies on building their own “private cloud”.

That makes a lot of sense if you are a company selling hardware and data centre software – that you want people to buy your stuff so it’s all about “private Cloud”. And the recurring message from Oracle (the ex-SUN guys) is that (1) Cloud is nothing new, (2) it’s just a logical progression of what you have been doing in your data centre, and (3) you can now just re-enginner your data centre and get to Cloud as the next step. I don’t happen to agree with any of those propositions, but that’s the sales pitch.

In a session on “Architecting the Modern Data Centre” Angus McDonald CTO of Oracle Australia/NZ explained the magnitude of the task. He placed a big emphasis on the challenges and the need for planning – even one years planning to face up to the challenges of energy efficiency, power ulitisation, green, on top of all the complex IT technical planning.

He talked about the typical low server utilisation – about 15% – space utilisation – about 30% – and power efficiency – about 1/2 to 1/3 of the optiumum – in the current generation of data centres. The need for consolidation before virtualisation, which he said often did not happen hence resulting in virtualising your mess. The need for simplifying, standardising, and automating your data centre.

You also have to throw in a software and data modernisation, which in Oracle’s world means you adopt their “Red Stack” from “disc to applications”. Fair enough, that’s what they are sellling. And this whole process takes place over 3 or 4 years, given that they recommend a year for finalising the data centre modernisation strategy alone.

So here’s the bit that I don’t get. At the end of this 3 or 4 years of yet another nightmare data center modernisation, you have what exactly – a data centre which may or may not approach the best practice and skills of a “Cloud” facility?

And you now start to REALLY invest in the training and retention of all your people to run it to Cloud standards. You’re not sure if you got the “green” bit right, or the power efficiencies, or the air conditioning, after all bigger things cropped up during the project. You’re not even sure if you got the virtualisation optimised as you gave up on “consolidation” since “servers are cheap”. You had to focus more on the self-service and the measurement/charging services, which turned out to be quite complicated. (Which is part of the lie about where corporate and government data centres at at today – they aren’t within a moonshot of allocating and charging for resources on a elastic basis so it is not just a “next step”!!)



And when it came to the converged infrastructure and merging the data networks and the storage networks that was a REAL distraction and you had to call in additional consultants as Oracle locally couldn’t come up with the answers you needed. But don’t say that Oracle didn’t warn you – they told you 4 years ago all the complexities and how much hard work and effort it would be. It’s just that you ran out of time and money to really do ALL the things that they recommended, so you cut a few corners.

But HEY! GREAT, now you are there.

You now have a data centre that is cloud-like, you’ve put in your $$$ millions, and you’re 4 years down the track. Now you can start on the application and data migration.

Now here is where I find myself out of sync with Oracle. This customer’s 4 year journey to a new data centre has placed them about 3 years behind what they could have bought from a public cloud provider all those 3 years ago.

Instead of building facilities, they could have spent the last three years learning how to migrate data and applications, and to manage the Cloud resources effectively to gain business benefit. This would seem the REAL “next logical step”, rather than getting buried in green, air conditioning, and energy efficiency issues for 4 years in order to end up 3 years behind.

Specifically, they could have invested a year in analyzing and implementing the minimum investment in the the current data centre to be able to synchronize their on-premise applications and data with a Cloud provider, and then over the next three years refine their Cloud versus on-premise architecture with a view to optimizing or eliminating the in-house TOC and footprint. And by the way Oracle’s Fusion Middleware could be a key part of a client’s solution for this on-premise off-premise loosely coupled architecture – or perhaps they would be better with a Blue or Green or Purple Stack?

To me, “Next Generation Data Centre” means something that you don’t want to own unless you are in the Cloud business. I certainly don’t think that it is simply a “logical next step” to do in-house, because in particular we are at a major transition in computing services which presents a whole new range of opportunities which did not exist even 3 years ago. The Cloud shift is the biggest shift in the last 20 years of IT and it’s certainly not business as usual ”let’s get out there and build our new data centre”!

What do you think of a 4 year journey to private cloud, for whom does it make sense?



Is there more business return in learning to sync with public cloud, or building a new data centre?

What size or what type of enterprise would you have to justify owning your own next-gen data center?

LEAVE A REPLY

Please enter your comment!
Please enter your name here