Abstract-High throughput is of particular interest in data center and HPC networks. Although myriad network topologies have been proposed, a broad head-to-head comparison across topologies and across traffic patterns is absent, and the right way to compare worst-case throughput performance is a subtle problem.In this paper, we develop a framework to benchmark the throughput of network topologies, using a two-pronged approach. First, we study performance on a variety of synthetic and experimentally-measured traffic matrices (TMs). Second, we show how to measure worst-case throughput by generating a near-worst-case TM for any given topology. We apply the framework to study the performance of these TMs in a wide range of network topologies, revealing insights into the performance of topologies with scaling, robustness of performance across TMs, and the effect of scattered workload placement. Our evaluation code is freely available.
An emerging architecture for software-defined data centers and WANs is the network fabric, where complex applicationsensitive functions are factored out, leaving the network itself to provide a simple, robust high-performance data delivery abstraction. This requires performing route optimization, in real time and across a diverse choice of paths. A large variety of techniques have been proposed to provide path diversity for network fabrics. But, running up against the constraint of forwarding table size, these proposals are topology-dependent, complex, and still only provide limited path choice which (we show) can impact performance.We propose a simple approach to realize the vision of a flexible, high-performance fabric: the network should expose every possible path, allowing a controller or edge device maximum choice. To this end, we observe that source routing can be encoded and processed compactly into a single field, even in large networks, with OpenFlow 1.3. We show that, in addition to the expected decrease in required forwarding table size, source routing supports optimal throughput performance, in some cases significantly higher than some past proposals. We thus believe source routing offers a clean abstraction and efficient implementation for future network fabrics.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.