Path Cost

Bridges use the concept of cost to evaluate how close they are to other bridges. 802.1D originally defined cost as 1000 Mbps divided by the bandwidth of the link in Mbps. For example, a 10BaseT link has a cost of 100 (1000/10), Fast Ethernet and FDDI use a cost of 10 (1000/100). This scheme has served the world well since Radia Perlman first began working on the protocol in 1983. However, with the rise of Gigabit Ethernet and OC-48 ATM (2.4 Gbps), a problem has come up because the cost is stored as an integer value that cannot carry fractional costs. For example, OC-48 ATM results in 1000 Mbps/2400 Mbps = .41667, an invalid cost value. One option is to use a cost of 1 for all links equal to or greater than 1 Gbps; however, this prevents STP from accurately choosing "the best path" in Gigabit networks.

As a solution to this dilemma, the IEEE has decided to modify cost to use a nonlinear scale. Table 6-1 lists the new cost values.

Table 6-1. STP Cost Values for Network Bridges

Bandwidth

STP Cost

4 Mbps

250

10 Mbps

100

16 Mbps

62

45 Mbps

39

100 Mbps

19

155 Mbps

14

622 Mbps

6

1 Gbps

4

10 Gbps

2

The values in Table 6-1 were carefully chosen so that the old and new schemes interoperate for the link speeds in common use today.

The key point to remember concerning STP cost values is that lower costs are better. Also keep in mind that Versions 1.X through 2.4 of the Catalyst 5000 NMP use the old, linear values, whereas version 3.1 and later use the newer values. All Catalyst 4000s and 6000s utilize the new values.

Was this article helpful?

0 0

Post a comment