The goal of this exercise is to analyze the community attributes observed in a RIB dump in order to gain experience with how communities are structured and how they are observed by collectors.
While the community path attribute
(RFC1997) is defined simply as sequence
of 32-bit numbers, each value in the sequence is encoded by placing an ASN in
the first 2 octets, leaving semantics of the second 2 octets to be defined by
the AS itself. As such, PyBGPStream makes the
community path available
as a sequence of dict objects, where each dict has the keys asn and
value.
In this exercise, you will modify the script that you wrote in Exercise 3 to:
-
Plot the CDF of unique ASes that set community attributes per peer (still using
route-views2). Consider only theasnfield of the community as the "key" in this case (i.e., build a per-peer set ofasnvalues). Why do different collectors see different numbers of ASes setting communities? -
Count the overall number of unique AS values seen in all community attributes for all peers of
route-views.sg. I.e., how many ASes do you observe setting at least one community value?. -
Plot the CDF of number of unique community
values per-communityasn. This will require building a per-AS set of observed community values.