On low-latency-capable topologies, and their impact on the design of intra-domain routing

Abstract

An ISP's customers increasingly demand delivery of their traffic without congestion and with low latency. The ISP's topology, routing, and traffic engineering, often over multiple paths, together determine congestion and latency within its backbone. We first consider how to measure a topology's capacity to route traffic without congestion and with low latency. We introduce low-latency path diversity (LLPD), a metric that captures a topology's flexibility to accommodate traffic on alternative low-latency paths. We explore to what extent 116 real backbone topologies can, regardless of routing system, keep latency low when demand exceeds the shortest path's capacity. We find, perhaps surprisingly, that topologies with good LLPD are precisely those where routing schemes struggle to achieve low latency without congestion. We examine why these schemes perform poorly, and offer an existence proof that a practical routing scheme can achieve a topology's potential for congestion-free, low-delay routing. Finally we examine implications for the design of backbone topologies amenable to achieving high capacity and low delay.

Tuneman: Customizing Networks to Guarantee Application Bandwidth and Latency

“…Homa [38] is a transport protocol for small messages in low-latency DCs, that assigns priority dynamically at a receiver and over-commits receiver's links to increase BW utilization. Low-latency routing is considered by [23] by dimensioning the physical topology. [32] is a measurement module for software switches.…”

Goal-directed graph construction using reinforcement learning

“…The proposed approach can be applied to other problems based on different definitions of robustness or considering fundamentally different objective functions such as efficiency [20], path diversity [54] and assortativity [55], which are of interest in various biological, communication and social networks. Since our formulation and algorithm are objective-agnostic, we expect that they are applicable out-of-the-box for other objectives, even those for which no strong baselines are currently known.…”

SCIONLAB: A Next-Generation Internet Testbed

“…Figure 6 illustrates the CDF of the path stretch for all the direct links in SCIONLAB. Over 60 % of the direct links result in a path stretch d a /d s ≤ 1.4, which is the threshold for an alternative to be considered useful [28]; this indicates that SCIONLAB users have alternative path choices that can be used without exorbitant delays. We also emphasize that approximately 20 % of the direct links score d a /d s ≤ 1.0 meaning that an alternative path with a better propagation delay exists.…”