Agility in IT Via the Cloud
Cloud Computing is all the rage today, leading some to dismiss it as just the next in a long line of fads which fade as quickly as they appear. But behind the buzzwords surrounding Cloud Computing there are sound best practices hard learned from years of struggling with day to day IT challenges. This post will outline those best practices, the technologies that enable and support them, and how they result in benefits immediately tangible to many types of businesses.
The traditional challenges that IT has faced are:
- Scaling: Every new piece of hardware adds additional dependencies and required maintenance, making scaling a service by simply throwing hardware at it a costly proposition.
- Reliability: As more and more of a business comes to rely, completely, upon the IT infrastructure in place, downtime quickly becomes a matter of serious money.
- Flexibility: Traditional technologies have encouraged and often forced IT departments to cast each particular piece of infrastructure into a specific role. This server is a mail server, that one is a file server, and so on. This freezes much of the infrastructure and adds a significant cost to re-purposing it as business needs change.
- Utilization: Industry estimates are that the majority of data centers today are seeing utilization rates of 10-15% — for every 100 hours the servers and infrastructure are running, they are idle 85-90 hours. Yet they still draw power and require rack space and cooling.
An IT organization cannot be agile until each of these challenges is mitigated by a set of techologies and/or best practices. One of the most important mitigating and game changing technologies of the last decade is virtualization.
It seems each new day brings more virtualization vendors to market. But VMWare and XenSource (now Citrix) have both pushed the state of the technology to the point that nearly every reasonably sized IT organization now has at least a virtualization pilot project in the works, if not fully deployed virtualized production environments.
Virtualization goes a long way toward addressing each of the traditional IT challenges listed above. Scaling a service, with proper engineering, is a matter of cloning more copies of the systems running a service. Reliability is increased because it is cheaper (in time, resources, and money) to spread a service across many virtual instances. Flexibility is gained when each of your physical platforms becomes a generic processing unit that can run one or more arbitrary virtual machines. And finally, increasing utilization of existing servers is inherent in the ability to deploy multiple virtual machines on to each existing physical server, netting the average data center at 10% utilization an 8x to 10x increase in processing capacity on existing resources.
These are the capabilities that Cloud computing brings. Once embraced, the infrastructure supporting a business becomes agile enough to respond to new and changing markets, while also better utilizing existing systems, increasing manageability, and, oh yeah, saving money along the way.