Coding Assisted Cloud-Edge Collaborative Computing

Cloud-edge collaborative computing has emerged as a promising solution to satisfy the demands for intensive computation and low latency. However, the straggler effect in cloud-edge collaborative computing systems is serious but has not been addressed. This can be mitigated by designing an effective computing scheme, with the assistance of cloud server. Moreover, the problem of transmission failure becomes more severe in cloud-edge collaborative computing systems. Specifically, the result downloaded from the edge nodes will be affected by poor channel conditions, while the result transmitted from the cloud to edge nodes will also suffer from packet loss. In this paper, we propose a coding assisted cloud-edge collaborative computing (CA-CECC) scheme to solve these problems. To make it more tractable, we separate the transmission phase from the computation phase, and decouple cooperative transmission from single-point transmission. By dividing the set of cooperative nodes into two sets, we further propose two simplified schemes of low complexity. Their latencies are established as upper bounds on the latency of CA-CECC. Numerical simulation results verify the superiority of our proposed scheme.