A Fundamental Tradeoff Among Storage, Computation, and Communication for Distributed Computing over Star Network

TL;DR

This paper characterizes the optimal tradeoff among storage, computation, and communication in a star network distributed computing framework, proposing a coding scheme that achieves Pareto-optimal performance.

Contribution

It introduces a new theoretical framework for the tradeoff analysis and proposes a coding scheme that attains the Pareto-optimal surface in star network topologies.

Findings

Achieves Pareto-optimal tradeoff surface.

Proposes a simple chain coding scheme at the access point.

Provides information-theoretic bounds matching the tradeoff surface.

Abstract

Coded distributed computing can alleviate the communication load by leveraging the redundant storage and computation resources with coding techniques in distributed computing. In this paper, we study a MapReduce-type distributed computing framework over star topological network, where all the workers exchange information through a common access point. The optimal tradeoff among the normalized number of the stored files (storage load), computed intermediate values (computation load) and transmitted bits in the uplink and downlink (communication loads) are characterized. A coded computing scheme is proposed to achieve the Pareto-optimal tradeoff surface, in which the access point only needs to perform simple chain coding between the signals it receives, and information-theretical bound matching the surface is also provided.