SDN: Software Defined Networks

Chapter 12. Use Cases for Input Traffic Monitoring, Classification, and Triggered Actions

Introduction

One category of use cases^[214] that seems to recur frequently is a variation on the theme of input traffic monitoring or classification, and then taking some sort of triggered action (or actions). The general premise is one of interception or detection of some traffic pattern somewhere in the network—often at the edge or access point—that then results in one or more triggered actions. The action or actions can vary and be quite robust: from as simple as dropping incoming packets or as complex as triggering a query to a radius server or an HTTP redirection. Once those actions are triggered, the system can either return to its original state and simply process traffic as if it had never happened, or alter its actions to do something else either implicitly, or as a consequence of receiving a response from a query such as a radius request. Let’s investigate a few canonical examples to help illustrate how this all might work, starting with the most basic input traffic interception mechanism available: the firewall.

The Firewall

At its heart, a firewall is a system comprised of an input traffic pattern-matching engine populated with a set of classification rules to match input traffic on. Classification rules range in capability from quite simplistic and primitive, to complex regular expressions. In all cases, each classification rule has a corresponding action that is taken by the engine ...