In case of huge data being processed - one can schedule several node calculations - those can be big ones in case of bigger data set is being processed. That can lead to temoprary overloading of CPU/memory access. One way how to resolve this and avoid trashing on OS level could be to allow suspend/resume of node calculation functionality.
I agree that this would be a very useful addition.
A slightly inelegant 'brute force' workaround, if your calculation is appropriate, is to run the calculation within a chunk loop, and use the loop end 'pause' to hold the calculation when required.