Support suspending InferenceService

[tenzen-y]

/kind feature

Describe the solution you'd like
[A clear and concise description of what you want to happen.]
Similar to batch/v1 Job, I would like to support suspending InferenceService.
The suspending semantics allows us to appropriately manage capacity planning in the hybrid cluster (mixed cluster of Training / Inference).

Also, the suspending feature allows us to manage all capacity allocation by Kueue.

Anything else you would like to add:
[Miscellaneous information that will assist in solving the issue.]

xref: kubernetes-sigs/kueue#1603

Links to the design documents:
[Optional, start with the short-form RFC template to outline your ideas and get early feedback.]
[Required, use the longer-form design doc template to specify and discuss your design in more detail]

[tenzen-y]

@yuzisun As we discussed in the last KubeCon, I would like to support Suspending semantics.
cc: @terrytangyuan I heard that you are interested in this feature in another place.

I will create the design doc, first.
/assign