Introduction

The amount of time an ISIS router takes to detect a failure can be tuned with some best practices, which relate to BFD, dampening, hello parameters or point to point timers.

Reducing Failure Detection

• The sooner networks can detect failure, the sooner they can react and propagate alternate routing information.
• Below are some of the features that help with detection:
  • BFD
  • IP Event Dampening
  • Hello Parameters
  • Point to Point Network type.

BFD

• Bidirectoin Forwarding Detection is an independent protocol that sends hellos on the physical interface.
• It has the publisher subscriber model, where ISIS subscribes to BFD events.
• Depending on the hardware platform, BFD can detect a failure in 100s of msec.

IP Event Dampening

• Prevents flapping interfaces from destabilizing the network of constant convergence.
• Using exponential delay, interface state is suppressed until interface is stable.

Hello Parameters

• Option should be used when BFD is not supported on an interface.
• Lowering the hello timers, increasing the bandwidth and CPU requirements.

Point to Point Network Type

• Eliminates the need to have a DIS or DR.
• No need to elect a new DIS.
• No need to flood CSNP for database synchronization.

Comments