berabera べらべら blahblah

This article presents a simple Logtilla log analysis module, log_geoip_stats, which gives the top N client countries, in terms of hits, from web access log files. This module uses the libgeoip-erlang library to get geolocations from clients' IP addresses.

I have recently released Logtilla, a framework for parsing and analysing web access log files, and showed in a previous article how to use Logtilla with a simple example. One of the most interesting aspects of Logtilla is in the inside. Logtilla demonstrates how to use ASN.1 to communicate between an Erlang program and a C port program. This article is both an introduction to the usual concepts of ASN.1, and a tutorial to apply ASN.1 to integrate Erlang and C programs. The protocol defined in Logtilla is simple, but the approach described can be used to implement sophisticated distributed protocols, especially appropriate for systems management applications.

I have written a small Erlang framework for parsing web access logs, called Logtilla, hosted on GitHub. This framework supports parsing logs in the Common Log Format, or in Apache's Combined Log Format. Thanks to the use of a C port program to do the parsing, Logtilla is very efficient: it can parse and analyze 15,000 entries/sec on my 4-year-old laptop.