There are principles to consider when managing workloads in a hybrid cloud model. Management, in this context, refers to how resources are allocated in order to handle workloads. Assignments may be based on resource availability, work priorities, or event scheduling.
In the era of unified central computing, workload management has been very straightforward. When a task had to be executed, a task was scheduled to run on that system. Instructions for running this task or function are usually written in a complex instruction language to control the functions. This set of commands helped the IT organization carefully plan the execution of workloads.
If a task-critical workload requires an inordinate amount of time to run, a set of instructions can be created to stop that workload and allow another workload to run. When the second workload has finished carrying out its task, the long-term workload can be resumed. If there are dependencies that a workload needs to complete a task, a command can be issued to search for that task so it can be executed and then add the result to the workload.
Keep the following principles in mind as you begin to think about managing workloads in a cloud model:
Understand processing requirements. You need to understand how your computing resources can execute your workloads on average and at peak demand. In general, IT often designs its computing resources to meet the maximum workload.
Use the modeling resources. You need to know the CPU, disk, and memory needed to perform the workloads. Generally, you create some kind of form to do this. Your model could be a simple linear model that calculates how much CPU each service has, or it could be something more complex.
Select the capacity you need. Optimize your resources based on required response time, number of services, and many other variables to consider depending on what you’re trying to achieve with your workload.
The challenge in managing any workload is to ensure that it can be carried out and delivered at the right level of performance. The principle is not that difficult if you are dealing with applications running on a single server. However, as IT infrastructures become more complex and heterogeneous (as in a hybrid cloud), doing so becomes more difficult.