Applications and Implications of a General Framework for Self-Stabilizing Overlay Networks

TL;DR

This paper presents a general framework for self-stabilizing overlay networks, analyzing its complexity and demonstrating its application to various topologies, including a newly designed topology for efficient convergence.

Contribution

It introduces a versatile framework for self-stabilizing overlay networks, analyzing its complexity, and designing a new topology optimized for convergence efficiency.

Findings

Framework applies to multiple topologies

Topology selection impacts convergence complexity

New topology improves convergence performance

Abstract

From data centers to IoT devices to Internet-based applications, overlay networks have become an important part of modern computing. Many of these overlay networks operate in fragile environments where processes are susceptible to faults which may perturb a node's state and the network topology. Self-stabilizing overlay networks have been proposed as one way to manage these faults, promising to build or restore a particular topology from any initial configuration or after the occurrence of any transient fault. To date there have been several self-stabilizing protocols designed for overlay networks. These protocols, however, are either focused on a single specific topology, or provide very inefficient solutions for a general set of overlay networks. In this paper, we analyze an existing algorithm and show it can be used as a general framework for building many other self-stabilizing overlay networks. Our analysis for time and space complexity depends upon several properties of the target topology itself, providing insight into how topology selection impacts the complexity of convergence. We then demonstrate the application of this framework by analyzing the complexity for several existing topologies. Next, using insights gained from our analysis, we present a new topology designed to provide efficient performance during convergence with the general framework. Our process demonstrates how the implications of our analysis help isolate the factors of interest to allow a network designer to select an appropriate topology for the problem requirements.