Cloud Gives IT an Off Switch
Traditional IT evolves in one way: bigger over time. Each time an application development project gets approved, IT adds more capacity, and once it’s bought, there’s no way to give it back. Servers, physical or virtual, remain online whether they are important to the business, or if they have long since lost their value.
When many users pool their capacity needs with a large public or private cloud provider, they can give back capacity they no longer need. Because the pool is big and configurations are consistent, the returned capacity will be useful to somebody else. Getting a bill for capacity consumed each month gives users a view of what they have, and a meaningful incentive to shut it off if it’s not important.
So when would IT consumers want an off switch? There are lots of situations where this would be useful, including:
1) Experimenting with speculative projects. Let’s say you’re a food products company, and you think that a social networking website might be a good way to engage with customers about products, compare recipes, share tips, etc. This type of effort could go either way - huge interest, global success, lots of increased revenue, or crickets chirping, users counted on one hand, basically a flop. If you have to buy dedicated hardware and estimate demand for the 3-5 years that hardware lasts, you’re likely to pass on such a project until you have a better guarantee of success, but if it’s cheap to start small and you can easily scale it up or give it back, you’re more likely to give the project a green light.
2) Manage spikes in demand. When IT assets are fixed in place for 5 years, you’re forced to build to accommodate the highest possible demand, even if you rarely if ever get close to that level. With cloud, you can increase capacity in real time as demand spikes, and then shut the extra capacity down when things settle. This allows you to pay less by not carrying excess capacity, and better address the spike by growing capacity as you measure the demand increasing.
3) Fine tune the configuration. When projects are built, a diverse team of business and technology people come together to define requirements and choose a configuration. And most of the time, they have little data on performance and usage patterns. With traditional IT, you make a guess about what the workload will need, and then live with that choice for the next 3-5 years. With cloud, you can start with your best guess, and then switch it out for a different one if you need more or less than you initially thought.
4) Deal with changing business realities. These days, mergers, acquisitions, new compliance regulations, and new lines of business are all par for the course. In a traditional model, these business-driven changes end up adding complexity and waste as IT struggles to adapt. With a cloud model, you can simply give back what doesn’t fit the new model, and scale up what does.
By Andrew Reichman
For more thought provoking cloud management insights visit vmware-erdos.com.