How to write a job scheduler for a distributed computing platform?

In the distributed computing system, in order to make efficient use of resources, we often need a reasonable scheduler to help us to accomplish the task's reasonable dispatch and operation automatically. Regardless of the system level or the application level. A well-designed scheduler is useful as long as the task is run on a system with limited resources.

On the operating system, we are in order to make full use of CPU, memory, disk, network. On distributed computing systems, we take full advantage of the abstract computational units of the Computing platform, excutor (threads, processes, memory, disks)

(To be Continued ...) ）

How to write a job scheduler for a distributed computing platform?