Packet Guide to Routing and Switching

Chapter 3. Spanning Tree and Rapid Spanning Tree

Ethernet structure and operation are well understood because the base protocol is consistent from one version to another and the standard behaves predictably in almost every topology. Since many of the decisions regarding Ethernet—such as the network interface, signaling, and equipment type—are pre-determined, one might say that Ethernet deployments are simple and straightforward. However, proper Ethernet network operation is also dependent on adherence to topology rules and other protocols, such as the address resolution protocol. So a simple network develops some interesting, and sometimes complex, characteristics.

This chapter is about the Spanning Tree Protocol and its faster version, the Rapid Spanning Tree Protocol. These protocols wage a continuing battle to prevent against loops in Ethernet networks. A loop in an Ethernet network is created when the topology is connected back to itself. This is a problem because unlike the Internet Protocol at Layer 3, Ethernet does not have any built in protection. Therefore, it cannot prevent frames from continuously circulating. When loops occur, user connectivity can be significantly degraded if not destroyed entirely.

The Spanning Tree Protocol is active by default, and is invisible to network administrators and users alike. But because it works and is “on” by default does not necessarily mean that we can ignore it. Sometimes spanning tree is very inefficient. “On by default” also means ...