Communications of the ACM
A Manifold View of Connectivity in the Private Backbone Networks of Hyperscalers
Credit: Getty Images
As hyperscalers such as Google, Microsoft, and Amazon play an increasingly important role in today's Internet, they are also capable of manipulating probe packets that traverse their privately owned and operated backbones. As a result, standard traceroute-based measurement techniques are no longer a reliable means for assessing network connectivity in these global-scale cloud provider infrastructures. In response to these developments, we present a new empirical approach for elucidating connectivity in these private backbone networks. Our approach relies on using only "lightweight" (i.e., simple, easily interpretable, and readily available) measurements, but requires applying "heavyweight" mathematical techniques for analyzing these measurements. In particular, we describe a new method that uses network latency measurements and relies on concepts from Riemannian geometry (i.e., Ricci curvature) to assess the characteristics of the connectivity fabric of a given network infrastructure. We complement this method with a visualization tool that generates a novel manifold view of a network's delay space. We demonstrate our approach by utilizing latency measurements from available vantage points and virtual machines running in datacenters of three large cloud providers to study different aspects of connectivity in their private backbones and show how our generated manifold views enable us to expose and visualize critical aspects of this connectivity.
1. Introduction
A salient feature of today's Internet is that large cloud and content providers, called "hyperscalers," are building and operating their own private global-scale infrastructures (e.g., Google,11 Amazon,4 Microsoft,9 Facebook,12 Akamai,13 and Alibaba1). These private network infrastructures serve to minimize the exposure of the traffic generated by their own suite of applications and services to the types of uncertainty and variability that their use of the "public" Internet (i.e., transit provider Internet) would entail. Accordingly, increasing portions of the overall Internet traffic utilize these large providers' private network infrastructures and thus bypass the public Internet.24
The growth and importance of these private infrastructures raise new questions, many of which call for empirical study. Historically, the principal tool used by researchers to understand Internet infrastructure has been traceroute. Developed in the late 1980s as a troubleshooting tool for network operators,10 the Internet measurement community has enthusiastically adopted and improved it to study the properties of the routes taken by packets and to infer connectivity in the underlying physical infrastructure as described, for example, in Spring et al.23
No entries found