The Cloud is Not a Product

Sunday, August 08, 2010

Rahul Neel Mani


Jim Whitehurst, President and CEO, Red Hat, in a conversation with Geetaj Channana, says that technology vendors must listen to their customers and not merely sell them their products...

Q:Have the platform wars, like the ones between Windows and Linux, turned into service oriented wars?

A: I think that platform wars are moving much more to a stack war. People are now looking for holistic environments to build and run applications. For most companies it is not about choosing between Unix or Linux or Windows — it's all the about the platform on which my developers will be developing — that's changing.

With Red Hat we have always been very operating system centric but now we have to think about reasonable well-integrated infrastructure components.

Q:Would you agree that vendors like you, Microsoft, Amazon and Google are trying to oversimplify the cloud in some manner?

A: Yes, because I do not think cloud is a product that a vendor brings to the table. My view of cloud, the way customers say cloud, and what they really mean, is a modern layered architecture where you have a choice of those various layers and application development capability that delivers rich quick application development. By definition, you cannot have cloud in a box. So, I do think there is a bit of an over-simplification.

Q:Is there a disconnect between what people are asking for and what is being delivered?

A: Absolutely. Most CIOs are not ready to move their applications outside of their data centre. There is so much hype around cloud; people are saying different things. But, when I go to CIOs, I hear three things –

1. They want to pay for what they use and when they use it – they don't like the model where they buy all the software and the hardware and take risks.

2. They want a layered modular architecture which allows them to move their applications and run it where it is best.

3. Richer, better and faster application development experience

People hear hype and they think that it is cloud – which does not necessarily mean cloud. And cloud done wrong actually makes these problems worse.

I think that is where the disconnect is, because when people say 'cloud' we literally take it as an off-premise solution to run functionality. I think that’s a mistake and is not really what the customer is asking for.

Q:You have said in your past interviews that PCs would die in the future – why and how?

A: PCs will exist but they will be a minority among devices because of the explosion of mobile. By 'mobile' I refer to mobile devices and mobile broadband that have changed radically the way we interact with technology. The iPhone, iPad, Android – those experiences are getting richer. The way we see the PC, may exist on some desk tops, but in terms of computing devices, it's going to have a smaller share everyday.

Q:Earlier the enterprise drove consumer technology – now, do you think the consumer is affecting the enterprise more?

A: Oh yes! When I talk to CIOs around the world, I hear over and over again that they are being forced to support consumerist things like software as a service or devices. And, they have to do it because everybody is demanding it.

People are going to blog, they are going to tweet, whether you like it or not. So, CIOs have had to set policies to help people connect their personal devices. I do think that it is a very different world where innovation is being led by certain Web 2.0 type companies and they are setting the pace for IT. And, it really puts a lot of pressure on the CIO – not only to support these devices and applications, but also puts a lot of pressure on the expectation levels of what they should deliver.

I would also like to talk about the death of version 2.0. There is no Google Docs 2.0 or 3.0, there is just rapid innovation that is adding feature and functionality. The whole programming model that is being used inside of large corporations has to change. And, that is as much behaviours as it is underlying technology. The usual way was to plan a new implementation and spend years spec-ing it and coding it - then many years later you introduce a new product. The pace of innovation and the way people work at home has radically altered the way you spec the work.

Q:Would you say the role of the CIO in an organisation has changed from being a guardian of technology to an enabler of business?

A: I think that the role of the CIO needs to change really to be an enabler of the business, but, I’ll be frank - half the CIOs are still struggling to keep the lights on, keeping close to the budget and just surviving - that they are not playing the role that they need to play. When 80% of your costs are lights on then you are always concerned about how you can keep those costs down.

Most CIOs need and want to be strategic but are completely hampered by their IT infrastructure. They have no time left whatsoever from keeping the lights on to focus on strategic parts of the business.

But IT drives business – it is a core of operation for many businesses, and presents new opportunities but in a day to day operation, the CIO is more concerned about how to keep their e-mail server running or how to keep their ERP going - things like that.

That is why we see an interest in the cloud - because of the frustration with what they have and what they have to maintain. They think that “If I could source it from somewhere else – from this miracle light shining on top of that hill” it would make life so much easier.

Q:But are people really doing that – shoving the old things out of the window and investing in the cloud - or are they just talking about it?

A: Oh, they are just talking about it. There are a lot of companies that are investing in development and testing on Amazon EC2 cloud or the IBM cloud. But, these are all development and test, lots of playing around. We also have some companies pooling and trying to set up private clouds but I would say that these things are in the advanced R&D stage now. And, most of it is because the CIO’s boss, the CEO or the CFO is asking them about it.

Q:Whenever I talk to the CIOs, they are sick of listening to the cloud!

A: This is because it is so far away from really being able to deliver on the components and the pain points like – flexible modular architecture, newer, faster ways to develop functionality and paying on what you use and when you use it.

We think that it is the cloud – but what is really happening is that the technology industry does not listen to the customer – we just take the technology that we have and shove it down the customer’s throat. So, cloud, from an IT industry perspective has turned out to be a hammer and everything else is a nail when actually, there are screws out there – we in the IT industry need a different tool.

That's what I say in Red Hat’s marketing too, if you want to pay by the drink – we do that already with our subscription model; you want a layered modular architecture – we do that already.

What we have to be careful about is what we understand from customers on what they think the cloud is – and so we are delivering off-premise computing. Off premise could be an option in the long term but what customers are demanding now is a solution.

This is where the disconnect is, and this is where I refer to the Gartner hype cycle – this is why there is going to be a valley of discontent.

Q:You entered pretty late into virtualisation space – how do you see yourself placed in the market within the next 3-5 years?

A: We see ourselves as one of the major players. This is because going forward, architecturally, virtualisation will be a part of the operating system. This is similar to the way 25 years ago you used to buy the TCP/IP stack for networking. Virtualisation is now a feature in the operating system it is not a stand-alone layer.

VMware was the first to find the market as a stand-alone hypervisor but there are so many architectural constraints to that. I think our model is superior. We made the switch in November 2009 - before that we were offering our version of Xen - and it was painful for us to explain to our large customers why we were making the change. But, it is really the next generation of virtualisation that is built into the operating system – you can get the benefit of reusing the code and security.

Secondly, you get the power of open source and Linux for hardware management. This what we announced at the summit, today with KVM and in the Fall when Red Hat 6 comes out, it will blow away ESX , whether it is in terms of performance, the number of cores supported, server density and other things. And I wish I could take credit for that, but frankly, that credit goes to Intel.

In five years our virtualisation share will be roughly the same as our operating system share. It will be difficult to say that for VMware though – when Microsoft and Red Hat both integrate virtualisation in our underlying operating system.

Q:This is a very interesting relationship that you share with VMware – you compete with them, yet you support that too.

A: Yes, it’s the same with Microsoft. We support Hyper on Red Hat Enterprise Linux(REL) and they support REL on Windows. Its an interesting set of relationships, but it is really important. They key to the next generation data centres is that they have to be layered. So we are adamant that we will neither lock layers nor provide the alternatives that our customers want at each layer. We have something called cloud foundation for that.

It is important for us because we cannot come out with a cloud in a box. Thus it is about giving your customers choice – at the hypervisor level, at the operating system level, at the messaging layer, application layer – all the way up the stack. We call it cloud foundation because it is foundational elements that you can take and set up together. They would help you build and manage cloud. That does not mean that it all has to be on the cloud – it could be on ESX or on Windows. Quite frankly, you cannot run ESX without buying their management suite. You are tied to the layer and stack together. The problem is that as soon as you tie them together, it is very difficult to move things around because you have to hardcode the links. On the other hand, if you have a nice clean layered model you do not face these problems.

However, this creates a problem for vendors as it is difficult to make money in these models because they are unable to lock-in their customers. A lot of these spaghetti architectures keep the cash registers ringing.

Q:Finally, how do you see Red Hat evolving over the next five years?

A: We are focused on defining the 21st century enterprise architecture. We may not deliver that architecture; we may deliver components of it, but not the whole. We are working hard to ensure that this architecture is defined by open layers that give customers a choice at each and every layer. Obviously, we do not supply every layer, (we are not in the database business) but we are working hard to create that architecture. This is a hypervisor that is open and inclusive, an operating system that can run on any hypervisor, an application server that can run on any operating system.

Cross-posted from CTO Forum

Possibly Related Articles:
Cloud Security
Cloud Computing
Post Rating I Like this!
The views expressed in this post are the opinions of the Infosec Island member that posted this content. Infosec Island is not responsible for the content or messaging of this post.

Unauthorized reproduction of this article (in part or in whole) is prohibited without the express written permission of Infosec Island and the Infosec Island member that posted this content--this includes using our RSS feed for any purpose other than personal use.