Re: tcp/ip rewrite for summer of code

[Pierre THIERRY]

Scribit Joshua Stratton dies 25/03/2008 hora 07:27:
> http://www.cs.utah.edu/~jstratto/soc_hurd/

You may under-estimate the difficulty of actually implementing the
various layers. But you seem to have done it before, and it's true that
reimplementing is far easier. But it seems your schedule lack any phase
of design, which would be critical.

You cannot just go and implement, you have to understand what the
tradeoffs will imply in terms of complexity, overhead and security,
especially as the Hurd is a somewhat different beast than classical
monolithic OSes. Many assumptions that you can rely on in a monolithic
OS don't hold anymore.

As far as security and efficiency is concerned, I highly recommend
reading 'Network Subsystems Reloaded: A High-Performance, Defensible
Network Subsystem'[1].

  1. http://www.cs.jhu.edu/~sarat/usenix-net-2004.ps

Having a capability-oriented network stack would be great, BTW.

Quickly,
Pierre
-- 
nowhere.man@levallois.eu.org
OpenPGP 0xD9D50D8A

signature.asc
Description: Digital signature