Rewind forward to 2009. Server virtualization changed the relationship between hardware and software and in the process has turned System Administrators into Capacity Planners for their shop. The planning for hardware capacity has changed as well. Unlike a mainframe data center made up of a few very large machines, a modern virtualized server room is made up of many more machines where VMs move from host to host often in automated manner making capacity planning a more real time exercise. Furthermore, unlike the mainframe where application workloads were predictable and static, in vitualized data center VMs are powered on unpredictably. Capacity needs for applications can change very quickly and become problematic causing performance bottlenecks. Just think what VDI will do in terms of rapid VM provisioning when hundreds of new VMs are launched at 9am when users come in the morning .
We have entered a new age of capacity management in which:
- Systems Admins are Capacity Planners
- Capacity planning must be on ongoing and tied to the rate of change in your data center
- Capacity availability must be monitored in close to real time
I welcome your comments
4 comments:
Alex, I agree. I get weekly questions like "do we have enough capacity to host project X?" or "will we need to add capacity before the end of the fiscal year?"
Having a good set of tools for performance monitoring & trending is key, as is knowing how to use them. It's also important to go slowly. Rather than provisioning hundreds of new VMs all at once, do a few at a time. That way, if you get into trouble, you can stop and figure it out. It's a marathon, not a sprint, right? :-)
At sxc.co.uk we are seeing a surge in companies establishing or re-establishing a proper capacity management function.
In a virtualised environment this is very important, and the cost savings (hardware, power, licenses. management) justify the purchase of a decent capacity planning and modeling application.
Alex, you have brought up a very interesting point concerning the optimization of resources. While the quarterly capacity planning is for all practical purposes not granular enough for the cloud, most if not all organizations can benefit from "Capacity Optimization". We used to call these two different functions that is capacity planning and performance tuning. The capacity part was planning for the CapEx cost of a new mainframe. The Performance tuning was stretching what resources we have to last (prioritize) until the next upgrade.
We utilized functions (Workload Management) like reducing the batch workload to less resources/priority during the day and then switching the online resources to less in the evening and letting the batch workloads have the priority.
I see the virtualized world needing more granularity as to which VM (i.e. workload) needs more resources than another VM.
There is still (despite some x86 Sysadmin's beliefs) a finite amount of resources available at any given time of the day.
History does repeat itself, if we only look through the right pair of glasses (perspective).
Cheers,
Bruce
I would like to comment on the original posting regarding CMG. The folks who belong to CMG - such as myself - have already embraced a more agile approach to performance and capacity management. These are no longer two distinct activities. They work together. Even with cloud computing, when you no longer have up-front capital expenses, the essential techniques used to track workload activity and growth and then map that to available resources has not really changed. The requirred agility and the technology domains have.
Post a Comment