Byzantine-Resilient Distributed Computation via Task Replication and Local Computations

Abstract

We study a distributed computation problem in the presence of Byzantine workers where a central node wishes to solve a task that is divided into independent sub-tasks, each of which needs to be solved correctly. The distributed computation is achieved by allocating the sub-task computation across workers with replication, as well as solving a small number of sub-tasks locally, which we wish to minimize due to it being expensive. For a general balanced job allocation, we propose a protocol that successfully solves for all sub-tasks using an optimal number of local computations under no communication constraints. Closed-form performance results are presented for cyclic allocations. Furthermore, we propose a modification to this protocol to improve communication efficiency without compromising on the amount of local computation.