Claude's Status Report - 01 of 13: Introduction, advancement, and proposition of an implementation

[Claude Hemberger <claude.hemberger () etu unistra fr>]

Hi everybody !

This is my first mail on this list, so I'll introduce myself first.

I'm Claude Hemberger, I am 22 years old, and I take part in the nmap
GSOC program.
I study computer science (more specifically networking and embedded
systems) and
I just finished the first year of my master.
My mentor is Henri Doreau.
I live in France (Alsace), and you can reach me under this address and
over irc
(sae@freenode)


The project that my mentor proposed to me is to parallelize the hostname
resolution. At the moment, only the reverse resolution is parallelized.
The goal of the project is to accelerate the scans with a big number of
hostnames as targets.

This first week has been rather slow for me, but here are my
accomplishments:

* Set up my working space:
  * Downloaded and compiled the sources
  * Created a branch in nmap-exp which I setup in git-svn
  * Read the code standarts and set-up a vim script that implements them
* Discussed the project some more with my mentor
  * The first meeting should be finished just as I send this email
* Read up on some documentation
  * skimmed over the book, and read up the parts relevant to my project
  * checked out other implementations of parallelized hostname resolution
    * nmap_dns.cc which does the paralellised reverse dns resolution
    * http://www.akkadia.org/drepper/asynchnl.pdf as recommended by my
mentor
* looked around the nmap source code, to find the interesting and
relevant bits
  * The most relevant part is line 373 in targets.cc, which is where
resolve is
    called on the hostnames. Resolve is part of the libnetutil.
* created a simple test to compare the speed of my version and the original
  version

Priorities:

* Think about the proposed plan of my mentor:
  1. Extend libnetutils to beautify nmap_dns.cc
  2. Change the resolve() implementation in order to use nsock
  3. add a parallelized hostname resolver
  (4. merge it with the existing reverse resolver in nmap_dns)
* Exchange ideas again with him
* Propose the planned changes to the dev mailing list
* Start coding :D



I'm open to suggestions or discussions.

Cheers,
Claude Hemberger
_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/