Every decision of investment into building a private cloud — whether it’s a private cloud for in-house use or a full-fledged cloud service provider platform (CSP) offering public cloud services — follows, in principle, a simple equation:
Investment and maintenance of the cloud < the return of investment over some time
Especially in these still-early days of publicly available cloud services, finding the correct number for the right side of the equation (the return on the investment) can be very vague and comes with a lot of uncertainty because the market for public cloud services hasn’t really started rolling yet. However, nobody would probably doubt that cloud is an important part of future IT and therefore cloud services will be needed in many various ways.
Another way to look at the equation is to try to lower the left side of the equation to make the decision easier from the financial perspective, for two reasons:
- First you have to get your business case approved in order to get the first cloud services started, so anything that lowers the investement part (the left side of the equation)—lowering the initial investment into a cloud environment—helps to lower the barrier (especially of your CFO) for getting the needed budget allocated.
- And as any installed system, after your cloud is up and running it will create costs — maintenance, license fees, electricity, labor costs, and so on. If you are one of the first movers, and your cloud services attract many clients, then you might be lucky not to have to look at the costs for running the cloud service platform — margins might probably still be quite good. But don’t worry, competition will catch up, and sooner or later you are going to face severe competition where you need to play the “price” card too,at least to some extent. If you haven’t kept this in mind from the very beginning when architecting and designing your cloud, you now might find it very difficult to change some of the underlying architecture and components you use. Long story short: Focusing on the costs of each component of your cloud environment — license and maintenance — will be essential also in the long run.
The Linux kernel-based virtual machine (KVM) is, to my mind, one of the first components that can help you to work on the the left part of the equation. Given that in bigger cloud environments, the number of virtual machines will easily reach thousands if not a magnitue of it, paying attention to choosing a cost-effective hypervisor can be one of those essential decisions you need to consider at the very beginning.
KVM has proven to be a stable hypervisor, easy to use and maintain and very cost effectiv – so much so, that IBM is also using it within its IBM SmartCloud Enterprise offering.
IBM SmartCloud Enterprise
When I started to dig for information about how IBM is dealing with KVM, I was quite astonished about the vast amount of information, links, and responses I received, in a short amount of time, from my colleagues. Many of them have already contributed large amounts of information in articles, blogs, and publications. Go to the following two links for more information about how IBM contributes, supports, and uses KVM—you might start to think too about taking advantage of KVM within your upcoming cloud architecture and design.
Virtualization@IBM (follow us on Twitter – @IBMVirt)
The Linux Library—featuring KVM