Monitoring Jitter in Software Defined Networks

TL;DR

This paper explores methods for monitoring end-to-end jitter in Software Defined Networks to improve Quality of Service for real-time applications, using controller-based and programmable data-plane approaches.

Contribution

It introduces two novel methods for estimating jitter patterns in SDNs, one leveraging a controller and the other utilizing P4 programmable data planes.

Findings

Simulation results demonstrate the effectiveness of both methods.

Controller-based approach accurately tracks jitter patterns.

P4-based method offers low-latency jitter monitoring.

Abstract

End-to-end jitter of a flow is an important metric that indicates the Quality of Service a user is experiencing, particularly for real-time applications such as video streaming, cloud gaming, and so on. Monitoring the jitter can help controllers make routing decisions for certain flows. This paper discusses methods in which we can estimate/follow important patterns in the end-to-end jitter of a flow. Two main approaches are taken, one using a Software Defined Network controller and the other using P4 programmable data-planes. Results from the simulations of each method are discussed.