Opinions
Server virtualization: clear thinking. straight talking.
Ian Marr – CO-FOUNDER and PARTNER
OVERVIEW: Ian Marr offers a point of view on enterprise virtualization. What actually are the benefits and why is it critical to approach virtualization from an application perspective? And why must you tackle complexity head-on?
Q: Ian, let’s begin by talking about CASE. In a nutshell, what is CASE?
Ian: CASE stands for Complex Applications and Server Exceptions. At the very simplest level, it is an approach developed by Virtual Clarity for improving P2V migration adoption rates and handling exceptions. CASE addresses applications that would otherwise block migrations.
Q: So is CASE a methodology that Virtual Clarity has developed?
Ian: I would describe it as more of an approach than a methodology. It is bespoke and appropriate to a specific client. We believe very strongly that there are no one-size-fits-all methodologies that will take the same form across different enterprises.
Q: So it’s all about flexibility, agility and client focus?
Ian: That’s right. CASE is aligned with the whole Virtual Clarity engagement model. We are very focused on the deliverable. We are not prescriptive about the process and we are not dogmatic regarding technology. In this sense, CASE fits in with our goal of helping visionary enterprises take intelligent steps on the journey towards virtualizing their entire infrastructure and transitioning to cloud computing.
Q: Perhaps we should begin by looking at the benefits CASE delivers. Can you give me a concrete example?
Ian: Let me answer that by referring to a specific CASE engagement with a major European bank. One of the primary benefits for this particular client was that they were able to close down datacentres. As a matter of fact, they managed to close them ahead of schedule. The task was all to do with maximizing the number of servers virtualized, but the objective was to do with emptying datacentres that were costing the bank a substantial sum each month to run and maintain.
Q: So CASE is not simply a smart approach to P2V?
Ian: That’s right. The CASE concept isn’t necessarily confined to P2V. It can be applied to datacentre and cloud migrations. If you want to migrate a datacentre, there are easy things to move out of the datacentre and there are hard things. Virtual Clarity’s focus tends to be on the hard things because we have wide and deep experience of solving these sorts of problems.
Q: So tell me more about what CASE does and doesn’t do.
Ian: CASE enables the acceleration of P2V – you can do more and keep doing more, faster. Bear in mind that CASE is about planning, in other words identifying what to do and giving advice. So the enterprise has to drive through to execution of those plans to derive the value. You then get the value from having a virtualized application.
Virtualization in action
Q: Let’s focus on some insights gleaned through client engagements. How did the client you’ve mentioned think about the process server virtualization?
Ian: At the outset this particular client’s approach to P2V was based on technology. From my previous experience at UBS, that was just not going to work. Out of their whole estate of Windows servers, the client wanted to kick off with a pilot to virtualize 300 of these over a 15-week period. The approach was one of ‘Let’s take all the Windows 2003 servers running software that we know we can virtualize.’ My opinion was that this did not represent the best way forward. The selection by technology resulted in a matrix of maybe 180 applications, across all the business lines and included random development servers, test servers and production servers from each application. It was simply not feasible in a 15-week period to talk to 180 application owners and none of the applications would have seen all of the benefits virtualization can deliver!
Q: So what advice did you give?
Ian: I advised the bank to look at these things from an application perspective. Because, once you approach the task in this way, you can go and talk with each business line and application, pick your priorities and go and virtualize all their servers on a rolling basis. Virtual Clarity also pulled together a ‘Lessons Learnt’ report for the bank; one of our principal findings was – ‘You’re leaving behind all the complex applications and all the hard stuff.’
Q: So what’s the problem?
Ian: Essentially, you come unstuck. You’re just carrying on blithely and ignoring the applications and servers that you can’t virtualize, so they end up as what you might describe as a steaming pile in the corner. If you want to try and close a datacentre, you’ve got to move absolutely everything out of it. There’s no point having one critical production server that people say they can’t virtualize. You can’t leave it behind because that means the datacentre can’t close. You’ve got to move it. For me, that’s the biggest justification for the CASE approach.
The value of virtualization
Q: Let’s return to the theme of the value of virtualization. So what is your personal view of that?
Ian: It’s not just about consolidation; it’s more about agility. The hardware abstraction is to do with being able to cycle the hardware at a different rate to the operation or the operating system and the application. So once you’ve abstracted yourself from the hardware, you don’t actually have to upgrade everything when you decide to refresh the hardware. You don’t have to upgrade the operating system, which forces an upgrade to the application. Instead, you can just do the hardware. All of a sudden, everything becomes less complicated and faster.
A second point of value is to do with availability. You can keep an application running and while you maintain the hardware, therefore the application is more available. If the application has no built-in high-availability features in terms of being able to cope with the failure of one of the machines, I can take a legacy application and put it on a virtual infrastructure. If it falls over, you can simply pick up that whole virtual machine and run it somewhere else. You don’t have to repair the piece of hardware – you can boot it somewhere else and repair the hardware in your own time. You can take that as far as running software within a full tolerant configuration where it’ll keep running if one machine fails – even legacy applications that couldn’t take advantage of that type of technology before. Virtualization enables it.
Q: What about disaster recovery?
Ian: Building on that, disaster recovery is so much easier. If one of your sites fails and you’re doing virtualization correctly at that level, you can simply shut that whole site or datacentre down and bring run it on the other datacentre.
You can easily maintain older versions of your applications, you can bring up a test environment or archive them for particular versions of software, you can share the hardware. You now have isolation between your test environments on the same physical hardware which you couldn’t do before. You can use your disaster recovery environment as your test environment. You can actually reduce your server footprint dramatically because you know that your production environment will run at your disaster recovery site because it’s exactly the same configuration. This looks like a consolidation argument, but actually it’s an extension of that.
The Cloud and organizational change »
The Cloud: CIOs as service brokers »
The Cloud: a newly flexible boundary »