Provide an in-memory StateBackend

[yahoNanJing]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

In some scenarios, there's no requirement of multiple schedulers. And when a scheduler is down, there's no requirement of task recovering and cluster topology recovering. In such cases, an in-memory StateBackend is enough and we don't need to persist the scheduler to the external storage. It's better to provide such kind of option.

Describe the solution you'd like

Describe alternatives you've considered

Additional context