Scheduling and Cluster Management

We have discussed some of the tools and methodologies used to provision our infrastructure; the virtual machines, networking, storage, and any other Azure services used by our applications. Now we need to schedule our services to run on the virtual machines that have been provisioned.

Scheduling is seemingly simple. For example, find and deploy five instances of my payment service in the cluster. However, this simple task can quickly become complicated as we introduce dependencies, constraints, optimizations, different types of resources, and requirements for those resources. A scheduler needs to take all these factors into consideration when determining the most optimal resources to deploy our services to.