How to Build Private CloudsBy Michael Vizard | Posted 2009-05-04 Email Print
The challenge: recasting internal IT operations into more of a service model.
One of the biggest strategic challenges facing IT organizations today is remaining competitive in a world full of cloud services that essentially outsource the IT function.
What makes cloud computing attractive to businesspeople is the concept of paying for IT resources as you go, as well as the flexibility associated with that computing model. In contrast, internal IT is seen as a fixed cost that executives have to pay for even if they are not utilizing all the assets at any given point.
The challenge for IT organizations is to figure out how to recast their internal IT operations into more of a service model that mimics the cloud computing model. The concept of managing IT as a service has been around for a while, but actually delivering IT as a service is much easier said than done.
But if internal IT organizations take a step back from the high-level rhetoric that routinely gets tossed around whenever cloud computing is discussed, it’s pretty clear that there are some fundamental new technologies
IT organizations need to understand and master:
Advanced Virtualization: If you think you’re on top of your game because you’ve deployed virtual machine software to increase your server utilization rates, think again. The need to provide dynamic sets of IT services means that you have to master the ability to rapidly provision an application stack on top of a virtual machine and then be able to manage it as it floats across multiple instances of virtual machines that need to be created as demand for that application grows on the fly.
Most IT organizations today do not have the tools or skills to manage virtual machine environments that, while being more flexible, actually introduce a lot more complexity into the data center environment.
Integrated Servers: In addition to acquiring new management tools, IT organizations also have to ask themselves if their existing data center infrastructure is really designed to handle infinite application scaling. Most internal IT organizations today run around silos of server, storage and networking equipment that need to be truly integrated into a holistic system. Provisioning of that equipment must be done in seconds and minutes in order to respond to rapidly changing business conditions, and because you can’t afford to throw labor at the problem, higher levels of IT automation have to be baked into the systems.
Scalable NAS: Just like servers, storage systems need to be able to scale on demand. That requires the capability to create pools of storage whereby one storage node is available immediately and additional available capacity is built into the architecture. The good news is that this also means the backup and recovery process, just as is the case with virtual machines, is automatically a part of the fundamental architecture.
Software Appliances: The core idea of this relatively new concept is that a preconfigured set of middleware is intelligent enough to look at the middleware infrastructure requirements of any given application and build the appropriate stack of supporting middleware instantaneously. No more waiting around for months while application development teams take inventory trying to figure out what the actual middleware requirements might be to make an application run.
Naturally, there are other technologies that will come into play. But if you want a fighting chance in terms of building a private cloud capable of delivering similar levels of flexibility and scalability to your organization, then these are the fundamental technologies you need to master.
Otherwise, most IT organizations are in danger of finding themselves becoming general contractors of IT services that are synthesized from various providers of cloud computing services. There’s no doubt that IT organizations will increasingly make use of those services, but the real cloud computing lesson that IT needs to learn is how to adapt the best concepts of cloud computing to the needs of the particular business it serves.
None of this is necessarily simple. However, it will ultimately lead to a higher quality of life for the people who manage next-generation technologies that were built with cloud computing concepts in mind, as opposed to those who are left struggling with trying to make the horse and buggy run better in a world where everybody else has a car equipped with a catalytic converter.