You might be under tremendous pressure to optimize your IT environments & perform an intelligent capacity planning to meet your organization’s future goals. While optimizing the environment, you would want to ensure that you have the right resource allocations as well as meet business service level or regulatory requirements. The policies in Capacity Planner allow you influence workload placement and workload sizing to ensure that the optimization plan conforms to business specific requirements.
Let's understand the different policy types supported by for capacity planning
1. Co-location & anti co-location
You specify two sets of virtual machines and specify whether these two sets must share the same host or must not share the same target host. In the anti-co-location case, no two virtual machines from different sets share the same host. In the co-location case, you want to put all of the set on the same host. If the virtual machines do not all fit on the same host, place the excess virtual machines on the next suitable target host.
For example, you would like to co-locate the workloads for database and web application server for an application on a single host. On the other hand, you would like to ensure that no hosts among competitor virtual machines placed in a shared cloud which can be taken care of by anti co-location parameter.
You would like to pick a set of virtual machines and specify them to be placed on an isolated set of containers, which is not shared by any other virtual machine that does not belong to that set. The container is a host or set of hosts. This would help you create an isolation zone for the specified set of virtual machines.
For example, you can create a Boundary of Oracle VMs. This ensures that all Oracle workloads are placed together on minimum number of target servers. This can help reducing licensing cost for software which are licensed per CPU core. Alternatively, you can also isolate certain workloads like to create a boundary for all WebSphere TEST VMs.
You would like to specify a growth factor for the processor resource demand of the virtual machines in excess of the current values of processor use that are displayed in the Virtual Machine. The growth factor can be for providing for future growth or preventing risk of resource overload.
For instance, you can tag important VMs as critical and add a policy “to use 20% CPU growth for all Critical VMs.”
You can decide to over-ride the sizing based on historic performance data for some critical VMs or would like to size them based on application best practices. This can help define custom sizing for workloads which are critical or for a specific application irrespective of past performance.
As an example, the Database administrators would recommend that DB2 servers having more than 1 Million transactions per second should have 8 GB Memory reservation.
Now the question is, do you really have to do all this manually. Absolutely no!!! IBM Smart Cloud Monitoring/IBM Tivoli Monitoring for Virtual Environments allows you to apply all these parameters as business policies while doing capacity planning or while optimizing your environment. Not only that, you can write your own custom business policies to cater other business compliances. For more information on this, refer IBM Smart Cloud Monitoring Information center