X
Tech

HP: Everything can be bladed

HP's head of blade strategy claims that the efficient and easy to deploy format has a big future beyond the server
Written by Colin Barker, Contributor

In the battle for the blade server market, it seems the gloves are off. HP announced the launch of a brand new blade system on Tuesday which it hopes will give it the edge in this increasingly competitive market.

Blades are one of the fastest growing sectors in IT. Offering plenty of computing power in a small format using industry standard Intel and AMD processors, blades offer IT managers simplicity and the possibility of making savings in areas like power and cooling.

The latest HP BladeSystem c-Class servers, according to HP, are packed with innovations that offer savings of 41 percent in the cost of server acquisition, 60 percent in data centre facilities costs and 96 percent in the initial set-up costs.

HP bases its claims on issues like cooling (the blades use low-power processors and an innovative, variable speed, temperature controlled thermo-logic system), power consumption (the boards have an environmental control system that minimises power consumption) and completely new systems management software.

To find out more, ZDNet UK talked to the man behind HP's blade strategy, vice-president and general manager for HP BladeSystem, Rick Becker.

Q: What’s different about the blade system your announcing today?
Becker: We have integrated a lot more into the architecture than the products we were delivering today. We are doing a lot to help customers measure power and cooling and once we can measure it, we can manage it. Once we can manage it, we can automate it. We now have instrumentation across the chassis and we have instrumentation in the blades.We know that many components are not manageable, but our belief is that once you instrument them, they will be.

Isn’t Insight Control (HP's system's management suite) already doing this?
Insight Control will enable our customers to set an energy envelope for the chassis or the rack automatically and bring in that resource as you need it. We will continue that innovation over the next twenty years. When I look at my R&D budget around power and cooling, automation and virtualisation we can do that.

So what about integration with business processes?
Well that’s Openview. That is absolutely a key component in this.

How does this fit in with the Adaptive Enterprise strategy?
We designed the BladeSystem c-Class from the ground up, knowing that as our customers deploy the Adaptive Enterprise that they were going to be doing so in a heterogeneous environment. They are not going to have “lights out”, automated computing across their whole IT deployment on day 1. So as we deliver the Adaptive Infrastructure, it has to be extensible.

So when we talk about Insight Control, that’s very specific to HP blade system but is integrated into the same console as Systems Insight Manager, Insight Manager manages blade systems, ProLiant and Integrity servers and StorageWorks - - it’s the only solution in the industry that is delivered on one console to manage your whole data centre.

It’s not as commonly known as it should be, but Systems Insight Manager manages any technology based on standards. So it they are using one of my competitor’s servers or storage devices … Insight Manager will manage that.
So today we already have a heterogeneous management console and what we are doing it delivering to that a brand new lights out capability.

So how does this stand against IBM?
Well I have shown you IBM’s blade servers today and I have shown you ours. It’s confusing to me how customers can deploy enterprise solutions without hot-plug drives, how they can support emerging technology with only four dimm, how they support a system that doesn’t have fully buffered dimm, and so on. My customers don’t have make those trades. And I have automated the processes. You can physically move from IBM blades to us at a mouse click.

You are going to do NonStop (HP's reliability engine) and storage in this chassis. When is that happening?
NonStop is not an announced product but we are announcing that we are moving NonStop to blades and we are doing the process pair on Integrity. We have publicly affirmed that we are going in that direction. I have Integrity blades in p-Class and I will have Integrity blades in c-Class this year.

So while it is not an announced product, I have architected an infrastructure, I have the blades and it is a stated direction. On storage I am announcing today the intent to deliver storage blades for direct attached storage this year.

What we do here is that we use the same components, the same processors, same memory, same storage across ProLiant, Integrity, StorageWorks and BladeSystem. So when I ask customers today whether they are going to networked storage or shared storage, almost all of them have a plan to go to shared storage at some time in the future but their business processes make it difficult for them to get there immediately.

By the end of the year I am going to have a storage blade so our customers will be able to take the same drives that they have populated today and move those to the storage blades, protecting their IT investments.

What this does for the system, the management and the middleware is it looks like one system with directly attached server storage. So when they are ready to go to a SAS or NAS environment they can just switch the blade. Wire once, change dynamically and deliver.

First half of next year, I am going to enable that dedicated storage blade to be a shared storage blade across the chassis so that any number of servers, inside the chassis, can act as one or more storage blades. To mix and match.
Now they want to be directly attached, I’m delivering that. Tomorrow they want to go to network storage in some fashion and I am making it very easy for them to pick when, where and how without ever having to throw away a single IT investment. Wire once, configure once, change it dynamically on the fly.

Why did you go to a 10u chassis that is not backwards compatible?
IBM picked backwards compatibility and it is not clear to me why they did that. If you took a look at the IBM blade centre and you use an old blade and a new edge, it does not take advantage of any of the new blade centre edge features. The only time you get the new features is when the new blades are in the new chassis. But because they kept backwards compatibility they don’s have enough dimms to be competitive, they don’t have hot-plug drives, etc. IBM is going to have to come out with a new generation architecture. They tried to fix the bugs they had in the old one instead of coming out with a new system. We took a much different approach.

We support our p-Class and it isn’t going away and it doesn’t have the problems that IBM has.

Don’t think that we are only going to do it in once chassis. 10u is what we are announcing today. When we look at our customers’ pain points it is not just about density or about budget. So when we took it from 6u to 10u it gives it much more air and power, so we could go to negative air pressure and have a manifold delivering Thermal Logic technology. We did that to give them more density in a real power envelope.

So customers are leading edge and have more efficiency in their data centres so we support up to 64 blades – the most density of anyone in the industry.

Will IBM copy you?
They are going to have to.

But IBM is already ahead of you in market share?
In Europe, I am number one with 41 percent of the market. Worldwide I am number two and IBM is number one by four points but I’ve grown faster in the last two quarters in revenue and units. It’s been a two horse race.

So you see blades replacing everything?
I believe everything can be bladed. So when we look across our applications, whether they be enterprise applications running HP/UX, Windows or SQL or Linux or Oracle, desktops or workstations there are going to be customers who leverage our BladeSystems and our Intelligent Infrastructure.

Now customers always want choice; so if you ask me if I believe that one day it will be all BaldeSystem and no rack systems? No, I don’t believe that.
But I do believe that we are going to continue to see rack systems grow – but by the way, blade systems is the fastest growing category – and we will continue to see large SMP systems like Superdome and Himalaya take advantage of our Adaptive Infrastructure and leverage blade systems in more and more places.

Now our model is different from anyone else’s because I leverage the design teams across the company, so the team that designs the ProLiant rack is the same team that designs the ProLiant blade and that is the same for Integrity and for StoargeWorks so you are always getting the latest innovations and HP is committed to choice.

Editorial standards