Mars: Near-Optimal Throughput with Shallow Buffers in Reconfigurable Datacenter Networks

TL;DR

This paper introduces Mars, a reconfigurable datacenter network topology that achieves near-optimal throughput with shallow buffers by emulating regular graphs, addressing buffer constraints that limit existing high-throughput designs.

Contribution

Mars is a novel periodic reconfigurable topology that emulates regular graphs to optimize throughput under buffer and delay constraints, unlike prior complete graph emulations.

Findings

Mars outperforms existing systems in throughput with bounded buffers.

Emulating regular graphs reduces buffer requirements compared to complete graphs.

Systematic analysis of degree d optimizes throughput given buffer and delay constraints.

Abstract

The performance of large-scale computing systems often critically depends on high-performance communication networks. Dynamically reconfigurable topologies, e.g., based on optical circuit switches, are emerging as an innovative new technology to deal with the explosive growth of datacenter traffic. Specifically, \emph{periodic} reconfigurable datacenter networks (RDCNs) such as RotorNet (SIGCOMM 2017), Opera (NSDI 2020) and Sirius (SIGCOMM 2020) have been shown to provide high throughput, by emulating a \emph{complete graph} through fast periodic circuit switch scheduling. However, to achieve such a high throughput, existing reconfigurable network designs pay a high price: in terms of potentially high delays, but also, as we show as a first contribution in this paper, in terms of the high buffer requirements. In particular, we show that under buffer constraints, emulating the high-throughput complete graph is infeasible at scale, and we uncover a spectrum of unvisited and attractive alternative RDCNs, which emulate regular graphs, but with lower node degree than the complete graph. We present Mars, a periodic reconfigurable topology which emulates a $d$-regular graph with near-optimal throughput. In particular, we systematically analyze how the degree~$d$ can be optimized for throughput given the available buffer and delay tolerance of the datacenter. We further show empirically that Mars achieves higher throughput compared to existing systems when buffer sizes are bounded.