Doug Clark, IBM UKI Cloud Computing Leader, in conversation with Sam Garforth (Part 2 of 3)

Part 2 of 3:  Doug discusses the changing role of IT

Read Part 1 if you missed it.

Sam: Is IBM PureFlex considered cloud?

Doug: Yes. It’s dynamic, it’s virtualized, it’s standardized, it’s got a lot of the attributes that you would use if you were defining cloud. It’s easy to deploy, it’s scalable, it’s a utility based environment. In this case, it’s probably a private on-premises cloud that you’d be installing, but then within that you would be able to cross-charge your various departments or lines of business or projects for the amount of consumption that they utilise. So it’s a private cloud and it would link into other clouds, external clouds, public clouds, multitenant clouds, and give you hybrid cloud.

Some might say it’s de-emphasising the role of IT, but it’s actually allowing IT to stop wasting its time and stop diverting its limited resource. Every business that I’ve ever spoken to is under some kind of resource constraint, and IT is no different. In many cases, the way IT has evolved over the recent decade with things like ERP and those kinds of extensions to the enterprise systems, IT has become more of a cost center and a support environment than it really deserves to be. As a result, a lot of the IT effort is dedicated to keeping the lights on, keeping these big grunty enterprise machines up to date, and patched and maintained; whereas with these new expert systems those resources can be pointed toward exciting new areas where the IT department will become a significant part of the business team in those organizations.

This whole concept of cloud enabling an agile environment is very true. The experiences we’ve had, the initial case studies we’ve got from our clients, suggest that by deploying cloud-based projects, you can do so much more and so rapidly, that you need to have the IT profession sitting right up at the front of the plane with the business. That to me is probably the most exciting thing in all of this because it’s repositioning a whole profession and putting it where it rightfully belongs, which is at the core of the business, not as a support function.  This is strategic. The reason why it has to be up front is that you can’t purely leave the responsibility to the business. We have heard of business users basically buying cloud services, almost using their corporate cards, just doing stuff on the fly. If IT isn’t involved, then there is no governance, no failover environments to support you for those processes if they become business critical. That could be very dangerous. So it is important that proper practices such as the IT Service Management (ITSM) Framework, ITIL, and the usual IT governance practices are included in these ‘front of plane’ environments.

Sam: Are ITIL and other governance processes current enough for the cloud world?

Doug: That’s a great question. In my opinion there’s an opportunity for practices such as ITIL to be challenged, tweaked, updated or re-evaluated in terms of whether they really fulfil the requirements that cloud forces on us. Are the current best practice review cycles and the current auditability sufficiently agile, or near enough to continuous to give you those guarantees that your cloud-centric systems are fit for purpose? Using the more historical type of environments and infrastructures, the changes were nowhere near as frequent or dynamic. What we’re talking about with cloud is almost a constant change so maybe those processes, which I still think are really valid, need to also become dynamic in their own implementation. All of that scar tissue that’s been built up over the years, all of that experience, you cannot undermine that.

There’s been a lot of negative hype around cloud, a lot of myths, all well documented in the public domain, where people have tried to short circuit or bypass IT best practices. Some of those examples have failed and cloud has been blamed for it. It’s not a cloud issue in itself, it’s predominantly a governance issue.

Sam: People say that governance slows down things. That cloud is actually not that much faster because we still have the business processes and approvals processes delaying things.

Doug: I have an analogy from industry of quality management. They moved from “quality control” to “quality built in.” Quality was built into the process with rigor from the beginning. It wasn’t a matter of a guy standing at the end of a production line saying “that one’s a dud” and “that one’s good” — this was building in quality. Why couldn’t we do something similar with IT best practice and build it in so it’s a continuum? Maybe you would still have audit-type snapshots where you have to stop and look at things in more detail, but in general it could be continuous.

Sam: How have you seen the cloud market change over time?

Doug: I talk to a lot of clients and one of the things that I’m finding fascinating, one of the things that really strikes me, is how disruptive cloud is. We’re saying cloud is disruptive from an IT point of view — yes, it’s changing the way IT is deployed. It’s changing the way IT is architected. It’s becoming more of a utility type model. Actually the architecture is very similar to a telco model with operations support systems that support network operation and management, service delivery and fulfilment, and business support systems that allow a dynamic environment,  pricing, billing, and so on. That’s different to how traditional business runs.

Something else that’s interesting is that the business model is disruptive because people, for example, are paying for workloads or “paying as you go” rather than buying physical kits or buying into long term contracts. There’s a school of thought that says long term contracts are now really under threat because cloud brings with it a fickleness in that it’s easy to get into and easy to get out of.  I dispute the latter but that’s one of the thoughts behind it.

The other thing that’s disruptive about cloud is the ecosystem. A lot of IBM Business Partners are becoming cloud service providers or managed service providers for the first time, so the center of gravity in the market is changing. We’re working very closely with a whole myriad of Business Partners in an ecosystem bringing in solutions, software, and offerings that will provide a service catalog.

Looking at the traditional innovation adoption curve, you used to be able to allocate a particular market, client, or industry and say they’re always a follower, or they’re an early adopter. You know from their cultures where they fit, but with cloud it’s changed. Industries that you think are going to be the first to move are often not, those that you think are normally laggards in the market are quite often the first adopters. It’s totally turning every dimension on its head. You can’t take anything for granted with cloud. It’s such a change that you can’t use any historic models to predict the future. Whether it’s an insurance company, a government agency, or a bank, or whether it’s a retailer, any of them could take the lead. This is a significant change agent and the benefits that those organizations are seeing from those first implementations of cloud are so significant that, if anything, it’s accelerating the curve not just switching it.

Sam: From a technical point of view, we’re moving to thin consoles and centralized processing, and people are saying that that’s what the future will be like. But these things are cyclical. That’s what it was like in the 1960s and then we went to client/server so is this going to be cyclical also? Will we go back; will it change again?

Doug: I think it’s a good observation. IBM has been doing cloud for 40 or so years, because mainframe is fundamentally a cloud-like environment. It’s virtualized, you can change workloads up and down, and provision and deprovision. So it’s not new to us in terms of how we tackle it from a technology point of view. But you’re right, we’ve seen a number of evolutions of client/server – moving workloads out to the periphery of the IT environment then reabsorbing it back into the center. I think with cloud, what you have now is a phenomenal choice and it’s really down to the IT professionals to select the most appropriate architecture. They need to determine whether this part of the architecture needs to be centralized because we want to pool capacity or CPU, because for example, it’s a high performance compute cloud requirement.  Another part of the architecture might need to be very distributed, because the point of presence is out on the edge, because that’s where the transactions are happening, because that’s where the real-time checking needs to be done, for example checking a passport ID or an RFID or a scanner.

We’ll end up with a blended architecture. Whether it’s centralized or distributed will depend on what’s fit for purpose. Just as we’ll end up with a hybrid cloud as the predominant result, I think we’ll end up with a hybrid IT environment, where you have some distributed workloads, because that’s best in class, and some centralized workloads, because that’s best in class. And those centralized ones might be multitenant. For example, we’re talking about high performance computing clouds for engineering environments, where multiple clients might buy into that capacity because they might use it on a project-by-project basis or they might use it seasonally and so the business justification is to share, rather than to own.