James Urquhart has been writing about the cloud for some time and in a recent post to launch his new gig at GigOm he began a process of trying to pull together some of his ideas from the last several years on the impact of the cloud on IT.
Urquart outlines a typical IT Ops workflow as follows:
- We buy a server
- We assign that server an IP address and wire it to a switch port
- We choose an operating system (which, I argue, is actually part of the server from an operations perspective), then install applications
- Finally, we monitor the health of the system based on–wait for it–server metrics: CPU and memory utilization, I/O rates, etc.
You can see where the subject near and dear to this blog comes into play in the last step where monitoring takes over.
Urquart went onto explain how these tasks map to a cloud environment, and I think it’s pretty clear how your role changes with the first steps as the cloud takes on these roles, but what’s not clear and what Urquart didn’t articulate clearly in this post (perhaps he will in a future one) was what happens to monitoring.
He breaks down the new world of cloud IT Ops into three broad areas: Applications Operations, Service Operations and Infrastructure Operations, but it’s not clear yet how monitoring fits into these new functions.
When your operations move to the cloud, one of the issues for IT Ops is just how you’re going to track and monitor the operation of these applications. As organizations move more critical operations into the cloud the need to monitor them only increases.
I wrote a post the other day on Amazon‘s first steps to bring EC2 instance monitoring into internal monitoring software via APIs. While the monitoring metrics are on the simple side, as I wrote it’s a positive first step in giving EC2 customers some insight into their EC2 instances from within the monitoring tools they are used to using.
So that last piece in Urquart’s list has to be more than a web site or an email alert when things change or go wrong with your cloud applications. You need to have more control than that, even in a cloud environment. And cloud vendors have to do more to acknowledge the needs of IT Ops in large enterprises.
The advantages of the cloud are well documented, and in many ways going to the cloud could free IT Ops to do more important tasks, beyond keeping things running, but you still need to be able to ensure the lights are on and the applications are working. Amazon’s announcement was important because it recognized this need.
While Amazon and other cloud vendors need to do more, when we outline the issues as clearly as Urquart has, we can begin to understand where the in-house IT Ops responsibilities map to the new cloud reality.
And that is also essential to understanding your roles as we make this transition and where monitoring fits in with all of this.