SOAR: Minimizing Network Utilization with Bounded In-network Computing

TL;DR

This paper introduces SOAR, an optimal algorithm for placing limited in-network computing devices in datacenter networks to significantly reduce network utilization, outperforming intuitive placement strategies.

Contribution

The paper formulates the problem of limited in-network device placement and provides an optimal, efficient solution for tree networks with arbitrary link rates.

Findings

Small fractions of in-network devices can greatly reduce network utilization.

The proposed algorithm outperforms intuitive placement strategies.

Performance gains are consistent across various workloads and network configurations.

Abstract

In-network computing via smart networking devices is a recent trend for modern datacenter networks. State-of-the-art switches with near line rate computing and aggregation capabilities are developed to enable, e.g., acceleration and better utilization for modern applications like big data analytics, and large-scale distributed and federated machine learning. We formulate and study the problem of activating a limited number of in-network computing devices within a network, aiming at reducing the overall network utilization for a given workload. Such limitations on the number of in-network computing elements per workload arise, e.g., in incremental upgrades of network infrastructure, and are also due to requiring specialized middleboxes, or FPGAs, that should support heterogeneous workloads, and multiple tenants. We present an optimal and efficient algorithm for placing such devices in tree networks with arbitrary link rates, and further evaluate our proposed solution in various scenarios and for various tasks. Our results show that having merely a small fraction of network devices support in-network aggregation can lead to a significant reduction in network utilization. Furthermore, we show that various intuitive strategies for performing such placements exhibit significantly inferior performance compared to our solution, for varying workloads, tasks, and link rates.