MANAGING RENDEZVOUS POINT REDUNDANCY IN A DYNAMIC FABRIC NETWORK ARCHITECTURE

摘要

Multi-destination frames in a network fabric may be carried in IP multicast packets. As such, the network fabric may us IP multicast technique such as a PIM protocol for handling the multi-destination frames. To provide redundancy, the system administrator can use phantom rendezvous points (RPs) that include multiple physical RPs where one of the RPs serves as a primary RP and the other RPs serve as secondary RPs (e.g., backup RPs). Instead of the system administrator manually configuring the phantom RPs, the RPs are automatically configured. To do so, the system administrator may use a GUI to provide multicast groups allocated for the multi-destination traffic, the number of desired phantom RPs (or physical RPs), and the desired RP redundancy. Based on these parameters, a data center manager generates one or more templates that automatically configure the network devices in the fabric as they are booted.

主权项

1. A method for generating one or more templates for configuring a network device in a network fabric, the method comprising:
receiving parameters for establishing phantom rendezvous points (RPs) in the network fabric; apportioning a plurality of multicast groups among the phantom RPs; establishing, based on the parameters, at least one primary RP and secondary RP for each of the phantom RPs using different values for respective subnet masks; and generating one or more templates based on the apportioning and establishing, wherein the one or more templates configure the network device to perform multicast packet routing when the network device is added to the network fabric.