Re: address pair exploration, flooding and state loss

[Iljitsch van Beijnum <iljitsch@muada.com>]

On 1-jun-2005, at 22:27, Erik Nordmark wrote:

> > AFAIK, applications that use raw sockets for actual user traffic  
> > are  extremely rare. (I can't think of any, not counting IP tunnels.)

> So what happens when DCCP starts being implemented, perhaps as  
> initially in user-level using the raw socket interface?

I'm not familiar with DCCP. Will it be implemented directly on top of  
IP? That won't be much fun in IPv4, with firewalls and NATs and all.  
Since general purpose transports need port numbers and a checksum  
anyway, running over UDP is generally a no-brainer.

> I am not comfortable assuming that the transport protocol or its  
> implementation must change in order for shim6 to be able to work,

I would agree about actually being able to work. However, the only  
thing we're talking about here is recovering from state loss. This is  
hardly a critical function, especially if we can make it work for 95%  
of all applications without any trouble.

> especially when there are alternative approaches which do not  
> require this.

Obviously an alternative approach that can do this without any other  
downsides would be great. But robbing a bit isn't easy.

> An approach where the shim failure detection can work more  
> efficiently when there is some optional advice from the ULP is  
> quite different than requiring that the ULP code change.

This is about state loss, not failure detection.

However, if we use the "let me know if you don't see traffic from me  
for N seconds" approach for failure detection, it could also detect  
state loss in many cases. (I.e., when the session isn't in idle state.)