Re: nse crypto

[Patrik Karlsson <patrik () cqure net>]

I've done some very/extremely basic tests that suggest that this won't
scale very well, and I don't think NSE is the problem. But like I said,
the tests are very basic, and as always I may be very wrong :)

Running the following command against a server listening on the loopback
interface takes roughly a second: ncat/ncat 127.0.0.1 9929 < /dev/zero

Running the same command ten times in a row, takes roughly 10 seconds.
Running ten times in a row in two separate terminal windows increases the
time for each sequence of ten connects to 20 seconds.

Three terminal windows result in 30 seconds....

//Patrik

Den 2011-03-13 12.28 skrev Toni Ruottu <toni.ruottu () iki fi>:

> Oh boy. I am lost. It does not use 100% cpu, it does certainly not use
> the full network bandwidth, and there is no way it could use up all
> memory. Now, what other bottle necks are there?
>
> On Sun, Mar 13, 2011 at 1:02 PM, Luis MartinGarcia.
> <luis.mgarc () gmail com> wrote:
> > On 03/13/2011 04:16 AM, Toni Ruottu wrote:
> > > Maybe the server only accepts one connection at a time from a given
> > > address?
> >
> > No, the server does not limit the number of connections in any way.
> >
> > Luis.
> _______________________________________________
> Sent through the nmap-dev mailing list
> http://cgi.insecure.org/mailman/listinfo/nmap-dev
> Archived at http://seclists.org/nmap-dev/


_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/