Q&A: How to Decide What Should Go to the Cloud
Reaching a decision to put some of your data and apps into the cloud is a no-brainer for most CIOs. For them, there are compelling arguments based on cost, time-to-market and access to move some operations out of the traditional bricks-and-mortar datacenter.
The tougher choices come when it’s time to decide what should leave the on-premise facility and where it should reside in the cloud.
"Our ability to deliver these IT services and make them more available and accessible to new markets has got to come with an economic balance of bringing them down to a way in which they can be delivered at the right price to the audience."
Zahl Limbuwala, the CEO and co-founder of Romonet, has been helping clients make those decisions on both sides of the Atlantic since 2006. Transform to Better Perform sat down with him to get his feelings on what is fueling the current transformation and what companies should look at before they begin their own transition to hybrid cloud systems.
Transform: Looking towards the future with the big data challenge we face, how do you see this data center transformation as it connects to business user demands?
Limbuwala: Basically, I see a couple of things. At the top level, I would say that it's the commoditization of the delivery of IT services that's having the biggest impact. So, we've gone from the use of IT as being a fairly specialist thing to IT becoming pretty much ubiquitous. So, everyone, everywhere, whether you're an adult or a child, whether you're at work or at school, or whatever you're doing, there's an element of IT services, and its applications in your life. That's pretty much the sole reason that we're having this explosion from a data perspective.
Transform: How does that “explosion” affect the delivery of IT?
Limbuwala: You've got a natural economic process based on a supply and demand. The demand is increasing, so the supply's increasing. Our ability to deliver these IT services and make them more available and accessible to new markets has got to come with an economic balance of bringing them down to a way in which they can be delivered at the right price to the audience.
You go from using multimillion-dollar enterprise software to being able to buy apps on your iPhone for 99 cents or even free. And there are different commercial models around them. So, there's that economic change which is driving the commoditization in the market, and that commoditization is being felt all the way through the technology supply chain.
Transform: OK, so how does that commoditization connect with the data center?
Limbuwala: If you look at the data center itself in terms of a box into which you put IT hardware and then you install software and so on, all areas of that technology stack are being commoditized. Go back 10 years, 15 years, and again, the IT equipment, the computers, the storage, the networking that was used there was very expensive, highly specialized, and required highly skilled people to look after it and maintain it. Today, we have gone from a world where if you filled up a rack in a data center with x86 equipment, you were talking about millions of dollars, you are now down to the tens of thousands of dollars. You've got lots more computing power, a lot more storage, a lot more bandwidth for a fraction of the price.
The assets inside my data centers have anything from an 18-month to a 3-year lifetime, whereas my real estate asset previously may well have been good for a full 20 years. That's a pretty fundamental change over a pretty short period of time, as well.
Transform: What should organizations think about as they move beyond that real estate model?
Limbuwala: The biggest problem for IT enterprise users is that they have very, very little visibility of cost within their environment. They understand that at a macro level, they're spending, whatever it may be, on their IT functions to provide the business, applications and business services that they use. But they don't really understand, or have any control or accountability over the decisions that get made on a day-to-day basis in their organization, and how those decisions impact the cost.
They can take fairly broad-brush estimations and say, "Well, I can save money by outsourcing my IT from my internal data center." And from total cost perspective, over a period of time, that may be true. It may not be true if I'm going to use that method of delivering IT services in a sustained way over a long period of time. And this is the challenge that they've got. It's understanding the business's requirements and then being able to make some sort of educated guess as the best way to fulfill that requirement.
Transform: Would you argue that the first step would be a hybrid model on the way to the cloud, depending on what your business needs are?
Limbuwala: Yes. I'd go a bit further and say that the corporate IT environment has basically been a ‘mono-culture’. As a user inside an enterprise, I had one choice for my IT. I went to the CIO and I said, "I need to host my server in a data center." Then the CIO said, "Well, we have 10 data centers we've built and we operate, and you can put it in any one of those. And the nominal price for that and the service level for that are essentially the same, because we've built them all the same."
What cloud has done is that it has broken that service mono-culture, and it's basically a large part of why we have this whole shadow IT situation now. It has allowed that user to bypass the CIO and say, "Actually, I don't need to put it in the corporate data center. I can just use my credit card and provision some servers on Amazon or Azure, and I can host it there." And the thing that that has had an impact on – again, is this initially obvious but not necessarily the case - is the realization that it's not just cheaper, it's quicker and easier as well.
Transform: Why is it cheaper, quicker and easier?
Limbuwala: Because the corporate IT guys says, "If you want a server, it's going to take me six weeks to order it and you've got to fill all these forms in, and then you're going to have to pay to own the server, and pay for the software to go on it," and so on and so on. Whereas in 10 minutes, I can pay with my credit card and provision the capacity that I need. I don't have to own the server if I don't want it.
Transform: What factors weigh against that approach?
Limbuwala: We all know the issues with public cloud and the concerns around security and data sovereignty and all these things, so I see a future really that looks like – I wouldn't necessarily describe it as hybrid, I guess, although it is. I'd see a future where as you look at each line of business and each application, you're able to make a decision. The business is able to make a decision that says, I have a number of ways of servicing this need from an IT perspective, and I have a number of factors that are going to influence it – data sovereignty, cost, speed, flexibility, time to market.
Transform: How do you see the transformation in North America and Western Europe versus other areas?
Limbuwala: I don't think there are any easy answers there. For example, where you're in a particular geography that's afforded clean, cheap power, the pressure on you to be much more efficient in the way that you utilize your IT capacity is very different than if you're in Germany, where it's 30 euro cents a kilowatt hour. And when you overlay that with, again, things like data protection and data sovereignty in Europe, it's a much more complex landscape. Because although you've got European level mandated data protection requirements, each country within Europe then implements that requirement as a minimum level. They can go over and above that requirement if they feel the need to.
Transform: What would you say to organizations that are hesitant to move their data into the cloud because of the privacy issues and the security issues?
Limbuwala: I don't think you can say it's advantageous to every business. There are definitely going to be businesses or organizations that are never going to be able to put anything in public clouds. Governments, security services, a whole raft of examples, that you would just never do it. And it may be because legislation doesn't allow you to do it. Obviously, you would never get the NSA hosting stuff on AWS. But if you take out any legislative reason, there are still instances where clouds won't necessarily make sense for your business. Again, I don't think it's ever going to be 100 percent in any particular direction.
I think people will have elements of their business that can and should absolutely be in the cloud. For example, email, nobody – big, small or otherwise, again, unless they're some sort of security services type organization – should be owning and operating their own email platform. They certainly shouldn't be doing it in their own data center. They should just be buying that as a service from some email cloud provider, whether that's Rackspace or Azure, or Office 365.
Transform: And that is something that your company does, correct? Analyzes these risks and the economics of it?
Limbuwala: Yes, we can tell you the costs and associated risks with any change in how you plan to provide IT. The risk is really identified by the customer. It's up to our customer to say, "Well, this is how important it is for me to be able to access email," or, "This is how important it is for this data to not leak out into the public." But with the risk identified, what we do is essentially tally the cost of the different options against each of those risks. With those risk ratings, we then put a cost against each of the different ways you could deploy your technology and your services."
CEO and Co-Founder
Point of view storyQ&A: George Rockett, DatacenterDynamics
Srini KoushikChief Technology Officer
Point of view storyThe Traditional Model of IT is Dead - Meet Exponential IT
Zahl LimbuwalaCEO and Co-Founder