 | Copyright © 2009, Juniper Networks, Inc. All rights reserved. Trademark Notice. | |
In many network topologies, point-to-multipoint (P2MP) label-switched paths (LSPs) are used to distribute multicast traffic over a virtual private network (VPN). When traffic engineering is added to the provider edge (PE) routers, a popular deployment option has been to use traffic-engineered P2MP LSPs at the origin PE. In these network deployments, the PE is a single point of failure. Network operators have previously provided redundancy by broadcasting duplicate streams of multicast traffic from multiple PEs, a practice which at least doubles the bandwidth required for each stream.
Ingress PE redundancy eliminates the bandwidth duplication requirement by configuring one or more ingress PEs as a group. Within a group, one PE is designated as the primary PE and one or more others become backup PEs for the configured traffic stream. The solution depends on a full mesh of point-to-point (P2P) LSPs among the primary and backup PEs. Also, you must configure a full set of P2MP LSPs at the backup PEs, even though these P2MP LSPs at the backup PEs are not sending any traffic or using any bandwidth. The P2P LSPs are configured with bidirectional forwarding detection (BFD). When BFD detects a failure on the primary PE, a new designated forwarder is elected for the stream.
For more information about traffic engineering, LSPs, and BFD, see the JUNOS MPLS Applications Configuration Guide.
This section discusses the following topics, which provide information about configuring ingress PE redundancy properties:
| Copyright © 2009, Juniper Networks, Inc. All rights reserved. Trademark Notice. | |