User's Manual

Chapter 12: Setting Switch Spanning Tree Protocols
138
Overview
Both STP and RSTP guard against the formation of loops in an Ethernet
network topology. A topology has a loop when two or more nodes can
transmit packets to each other over more than one data path. Packets can
become caught in repeating cycles, referred to as broadcast storms, that
needlessly consume network bandwidth and that can significantly reduce
network performance.
STP and RSTP prevent loops from forming by ensuring that only one path
exists between the end nodes in your network. Where multiple paths exist,
these protocols place the extra paths in a standby or blocking mode. In
addition, STP and RSTP can activate redundant paths if primary paths go
down. These protocols guard against multiple links between segments
and the risk of broadcast storms and maintain network connectivity by
activating backup redundant paths.
One of the primary differences between the two protocols is in the time
each takes to complete the process referred to as convergence. When a
change is made to the network topology, such as the addition of a new
bridge, a spanning tree protocol determines whether there are redundant
paths that must be blocked to prevent data loops, or activated to maintain
communications between the various network segments. This is the
process of convergence.
With STP, convergence can take up to a minute to complete in a large
network. This can result in the loss of communication between various
parts of the network during the convergence process, and the subsequent
lost of data packets.
RSTP is much faster than STP. It can complete a convergence in
seconds, and so greatly diminish the possible impact the process can
have on your network. Only one spanning tree can be active on the switch
at a time. The default setting is RSTP.