Charles Babcock, author of the book “Management Strategies for the Cloud Revolution”, offers some great insights on cloud computing. Consider these seven things before converting to cloud computing.
First of all, he gives the notion a proper definition, as every vendor describes it differently. In its simplest form, cloud computing is a “new method of distributing computing power,” says Babcock. All the pieces that make it up have been around for a while, but it’s the combination of high bandwidth and low-cost cluster servers makes cloud computing practical (and affordable, as it’s on a pay-as-you-go scheme).
Cloud computing breaks you of the bad habit of over-provisioning your own data center, that is, applying a new server for every app that you need or use. Even if you’re committed to keeping your own servers intact, you could use the cloud to handle overflow and conserve capacity internally.
Nobody ever said you “have to” put all your data on the cloud. After all, some of your data is sensitive and will always remain so, and you want to keep it as private as possible. So, you may choose to keep that information on internal servers. Other data that you elect to put on the cloud should be tracked and monitored constantly for security, compliance and potential abuse.
Consider data and workload replication because you get excellent disaster recovery support and the juice you need to cover transaction surges.
Make sure you know the details of your workload before searching for a cloud provider, including CPU types, average workload and capacity, and whether the workload is storage or transaction intensive. Share this information with your cloud partner so that you can both agree, as in an SLA, how they should handle (or not) your regulated data.
Chances are you’ll wind up with a mix of public and private clouds. The way things stand now, most corporate data centers employ servers that are partially virtualized and are on their way to becoming private clouds. You should create a variety of server templates for your workload, as many as necessary, and ask your cloud computing provider to replicate them in their data center.
Shop for a partner that can accept your templates. That will make it simpler to roll your servers out into their cloud support structure.