BGPStream Lab, TMA PhD School 2017

Lab Prerequisite: BGPStream and PyBGPStream

Please ensure that you have BGPStream installed prior to attending the lab. Contact bgpstream-info@caida.org with any problems that you encounter.

For this lab, you will need to have access to a machine with BGPStream and PyBGPStream installed. This could be on your laptop that you will have with you in the lab, or on a remote server that you will access e.g., via SSH.

The BGPStream website has install instructions for a variety of operating systems.

Code Organization

You should clone this repository and use the existing exercise directories to organize your code -- this will also give you something to refer back to in the future when you want to make use of BGPStream.

Exercises

The first two exercises are simple and designed to get you familiar with using BGPStream, first from the command line (using BGPReader), and then programatically from within Python (using PyBGPStream).

• Exercise 1: Using BGPReader
• Exercise 2: Using PyBGPStream

The remaining exercises all use PyBGPStream and incrementally build on the code you wrote in Exercise 2. While each exercise builds on the code written in the previous exercise, we strongly recommend you work on each exercise in a separate file.

• Exercise 3: Peer routing table sizes
• Exercise 4: Communities
• Exercise 5: AS path sanitization
• Exercise 6: Triggering active measurements

Troubleshooting

If you plan to run BGPStream on your laptop (or a machine with limited RAM), you may want to run the following in your shell before running bgpreader or any scripts that use PyBGPStream:

export LIBTRACEIO=nothreads

This will force BGPStream to use a single thread for reading BGP data, and will use significantly less memory.

Useful Web Services

• Convert between human dates and epoch seconds (since BGPStream currently only understands seconds since the UNIX epoch): https://www.epochconverter.com/

• Look up prefixes and ASNs by name, look up prefixes by origin AS, etc.: http://bgp.he.net/

• Look up all kinds of interesting things about Internet resources (IP addresses, ASes, Prefixes, DNS names, etc.): https://stat.ripe.net/

Helpful Code

Elem generator

You may find it more convenient to define a generator function that hides the code to extract Records and Elems from the stream. The next version of PyBGPStream will include an API similar to this.

First, define the generator function (you can use this code verbatim):

def elem_generator(_stream):
    _rec = _pybgpstream.BGPRecord()
    while _stream.get_next_record(_rec):
        while True:
            _elem = _rec.get_next_elem()
            if _elem is None:
                return
            yield (_rec, _elem)

Then, instantiate the stream like normal and use the standard Python for ... in ... syntax to iterate over records and elems in the stream:

bs = _pybgpstream.BGPStream()
# configure the stream here
bs.start()

for (rec, elem) in elem_generator(bs):
    print "%s: %s" % (rec.collector, elem.peer_asn)