Storage, Computation, and Communication: A Fundamental Tradeoff in Distributed Computing
We consider a MapReduce-like distributed computing system. We derive a lower bound on the optimal communication load for any given storage space and computation load. This lower bound matches the achievable bound we proposed recently. As a result, we completely characterize the optimal tradeoff between the storage space, the computation load, and the communication load. Our result generalizes the previous one by Li et al. to also account for the number of computed intermediate values.
READ FULL TEXT