Given that HA clusters can reduce downtime, and virtual machines are portable, the next logical question is, why spend the money for a company server? After all, properly hosting a server requires non-trivial environmental investments around space, good power, cooling, and operations. Ideally, the server should sit somewhere isolated from significant human traffic, in an area that stays comfortably cool, and that has an uninteruptable power supply to provide clean power. Somebody needs to back it up and maintain a rotation of offsite media. Somebody needs to keep it up to date.
What if the server were really a virtual machine located in an HA cluster at a commercial hosting site? And what if the desktop computers were also virtual machines residing in that same data center? Theoretically, a business could operate this way, outsourcing all the capital cost and expensive environmental factors to a hosting service with economies of scale. This is the promise of clouds, and the technology works, but the industry has some significant security and other hurdles to overcome before this becomes feasible on a large scale.
Meantime, clouds can play a significant role right away to handle demand peaks. Consider a florist with a holiday such as Valentines Day or Mother’s Day approaching. Volume might double or triple during the few days leading up to these holidays, then remain constant during other times. Instead of buying a system large enough to handle that peak load and keeping it idle the rest of the time, what if that florist could design its applications to harness capacity in the cloud to service those peak demand periods, but keep normal operations in-house?
Clouds might be a great option here. But before taking the plunge, here is a more detailed discussion.
Contact us and we can help sort through the options.