MBGP RPF

Data collection's purpose: to monitor the number of sources in each "half" of the Internet multicast infrastructure - PIM-over-MBGP and DVMRP-plus-PIM-over-DVMRP.

How the data collection works:

• Collect ipMRouteTable from fixw-mbone.nsn.nasa.gov via SNMP
• Use ipMRouteOctets to identify entries that have been used to forward packets.
• Use ipMRouteRtProto to identify what protocol was used to create each entry.

The data is collected every 5 minutes. Because of the high volatility, I'd like to collect it more often but using SNMP it takes a couple of minutes to collect the data. (The high volatility is not solely due to entries which have not been used to forward any packets; see the different graph styles below.)

This metric is flawed in that the number of sources in the whole infrastructure is dependent on what multicast events are occurring, so no conclusions can be drawn from the absolute numbers. In addition, the number of DVMRP entries is dependent upon the state of fixw-mbone.nsn.nasa.gov's conneciton to the DVMRP infrastructure, which is unreliable (see relevant data collection).

There are five types of graphs:

1. Stacked, real first then 0-forward entries. This graph represents entries that have been used to forward traffic as stacked shaded areas at the bottom of the graph, and entries used to forward no traffic as lines stacked on top.
2. Stacked, by protocol. This graph represents all values as stacked shaded areas.
3. Stacked, no 0-forward entries. This graph simply contains the stacked shaded areas from graph #1, without the 0-forward entries to make it easier to read.
4. Non-stacked. This graph simply has a line per protocol, so that variations in one protocol will not affect the apparent value of another.
5. Percentage. This graph is similar to graph #4, except that the total number of (real) entries is scaled to 100%.

There is no data available on entries used to forward no traffic before September 1999. There is no data available for the period February 1999-August 1999; this represented a time during which the data collection was suspended (the router being collected from switched authentication mechanisms and it took me a long time to get the SNMP data collection working).

This material is based upon work supported by the National Science Foundation under Grant No. 9729498. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation (NSF).