asn_report is a small Flask application that parses traffic destined to the internet and displays a chart of to get an idea of which ASes you send more to.

It's composed of two things:

1. Python utility that collects or captures traffic, parses, and stores into database. Scapy is used for live capture and currenly netflow v5 is supported for collecting.
2. Flask web application that reads from the database and creates a couple charts in javascript with the help of chartkick.

asn_report is very easy to install. Just make sure you have git installed and:

git clone https://github.com/coxley/asn_report
pip install -r requirements.txt .

Usage is pretty straightforward. There are two binary files that will be installed into your environment.

To start storing traffic info into database, follow the usage doc for asn_capture:

asn_capture

Description: Performs AS lookup for the destination of every packet processed
             and stores in database for later viewing in web summary.

             Traffic can either be gathered via a live capture with custom
             filter (pcap) or via netflow by sending flows to the process
             (flow).

Usage:       asn_capture (pcap|flow) [options]

Options:

    --help                          Show usage
    -f --filter=<pcap_filter>       PCAP filter syntax. Passed directly to
                                    scapy. [default: ip]
    --nflow=<flow_version>          Version of netflow to collect. Currently
                                    only 'v5' is supported. [default: v5]
    -h --host=<host>                IP address to bind flow collector on
                                    [default: 0.0.0.0]
    -p --port=<port>                Port to listen on for flow collector
                                    [default: 2303]

Capture should be started before webserver because it will initialize the db if it doesn't exist yet. If using netflow, once started feel free to start pushing flows to it from your router, fw, etc!

To start the webserver to easily look at summary of the data:

asn_report

That should spawn a local webserver reachable via http://localhost:5000/

For the graphing, you have the option to display each AS the following ways:

1. AS[num]. This is the default way.
2. AS[num]: [owner_string]. This is what I called 'display name'. To enable this provide --display-name as the only argument to asn_report
3. [owner_string]. I assumed that the display name could get pretty long so this final option is if you just want the name to show without the AS. To enable it, provide --owner as the only argument to asn_report

Only Netflow v5 is supported because I haven't been able to find a v9, sflow, or ipfix collector written in Python. Netflow v5 is pretty simple and I was able to find a few examples already written to tweak. Who knows, maybe at some point I'll fill the gap.

--

There are two offline databases that may need updating from time to time. These were chose to greatly speed up lookups and not to bog down other people's services.

First is the database used by the pyasn module. This is a converted version of a RiB snapshot taken from a looking glass. It's about 12MB and maps AS to IP prefixes.

Second is the MaxMind Organizational CSV. This one maps AS to Org names and is about 12MB as well unzipped.

To update these files, the easiest way is to just reinstall by grabbing the repo again and making sure you have git install as well as unzip. Then before installing with pip, run the update-databases.sh script. This will download and convert/extract both databases and then add them to git repo so setuptools notices them.

--

The main object asn_report is a custom class ASNLookup that glues together a lot of these for easy back and forth.

As of right now, the DB schema looks like: