Re: [GSOC] Nmap exclude list implementation

[Harshil Lodhi <lodhi.harshil () gmail com>]

I am starting to profile nmap for exclude list implementation. I wanted to
know does the order in which the addresses are stored in the exlude list
file ,matter in the lookup. I mean if they are sorted, does it speed up the
process. Does anyone know after what size of the exclude list, does it
start to slow down.

On Wed, Mar 4, 2015 at 3:27 AM Harshil Lodhi <lodhi.harshil () gmail com>
wrote:

> Hi Jacek,
> Definitely. Performance measurement is a must before optimizing random
> parts. I always remember Donald Knuth's quote  "Premature optimization is
> the root of all evil". From the experience that I have in programming in
> C/C++, these advanced data structures are useful only when the size of data
> becomes big. A normal array/ linkedlist would be fast for a small sized
> list but when the size becomes big, the asymptotic nature starts to
> dominate. I first plan to profile the nmap using gprof to see the exact
> part of the code that is taking time and then can decide on the
> optimizations.
> The problem is mentioned in the gsoc page of Nmap (http://nmap.org/soc/)
> . I discussed its implementation with "bonsaiviking" on irc and came to
> this conclusion. I first plan to profile nmap to see.
> Also I would love to work on more optimizations for Nmap apart from this
> if need exists.
>
>
>
> On Wed, Mar 4, 2015 at 3:08 AM Jacek Wielemborek <d33tah () gmail com> wrote:
>
> > W dniu 03.03.2015 o 20:00, Harshil Lodhi pisze:
> > > Hi everyone,
> > >
> > > I was going through the GSOC ideas "Performance/Optimization Specialist"
> > > one. In their its mentioned that current implementation of exclude list
> > is
> > > not a good one. It takes order of the size of the list. David
> > particularly
> > > mentioned about implementing the lookup using BDD.
> > >
> > >
> > > The link for the paper doesn't seem to work for me. I searched for the
> > > topic and read about it. One of the alternatives that we can consider is
> > > Patricia/Radix trees. They seemed to be more widely used than BDDs.
> > > Following is a highly cited research paper for the same.
> > > http://ece.ut.ac.ir/classpages/F83/Advanced%20Computer%
> > 20Networks/PAPERS/LOOKUP/routing.pdf
> > > I am very much interested in this position and in the script developer
> > > position. I have been using the proxy module of Nmap for the past 3
> > years
> > > to find working proxies inside my university campus during nighttime
> > when
> > > the main internet is shut down.
> > >
> > > Waiting for your feedback on this.
> > >
> > >
> > >
> > > _______________________________________________
> > > Sent through the dev mailing list
> > > https://nmap.org/mailman/listinfo/dev
> > > Archived at http://seclists.org/nmap-dev/
> >
> > Hello,
> >
> > I like this idea, but just to make sure - how did you pick this
> > particular problem? I'm asking to make sure you plan to make some
> > performance measurements before you start optimising random parts of
> > Nmap functionality.
> >
> > Cheers,
> > Jacek
_______________________________________________
Sent through the dev mailing list
https://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/