We talk so much about provisioning. There is a lot of complexity there, and there’s a certain magic to creating cloud service from component parts. It’s interesting, and the possibilities are endless. It may be the most compelling piece of the cloud story. Magic in moments.
Any IT administrator, datacenter architect, or VP of operations will tell you that provisioning, while fascinating, is a point-in-time activity. Operations is forever. And operations includes, as David Chernicoff notes in his blog, monitoring of the cloud services to ensure they are meeting SLAs. I couldn’t agree more. I would also add that it includes:
- Patching the cloud services
- Managing capacity in the environment
- Ensuring both configuration and operational compliance
- Managing the finances, to the extent that chargeback or showback are implemented
- Change management and CMDB integration
- and all the rest of those great things that have kept IT environments running smoothly for years
Users don’t see these things. They see services running slow, compatibility problems, security alerts, and all manner of errors. When there are no errors, they assume nothing is happening – because, why would it? All is well. And, while we’d all love a gold star for our heroic efforts to keep all systems normal, we also innately know that the freedom to be ignorant of the ongoing operations is one major value of cloud services to our users.
And with hundreds of cloud services humming along on the same shared infrastructure, changing and moving and de-provisioning and being created, keeping the lights on is no small task. Ironically, that’s where the true magic happens.