Re: udp retrieval protocol

Quoting Lester Vecsey <[email protected]>:

> From :http://cr.yp.to/im2000.html:
> 
> --
> How should messages be identified? How should messages be downloaded?
> Messages could be retrieved through HTTP, but an NFS/FSP-style UDP-based
> protocol would be much more resistant to denial of service.
> --
> 
> 
> A few of us would come up with these test servers and put some messages in
> them, and write clients to talk to them, basically a set of unix utils that
> talk to the server, 

I have recently gained access to a UML (user-mode Linux) hosting service and 
would be willing to donate a portion of its capacity/bandwidth for a testbed. 
For those of you not familiar with UML, it is basically a kernel that runs in 
user space, and allows multiple incarnations on a single physical box 
(frequently used in honeypot traps). We can attempt to simulate traffic flows 
between several virtual mailstores, even implement different protocols to 
evaluate which are the common features (which goes into a RFC) and which can be 
specific implementations (for those inclined to wander down the 
commercialisation path).

The only sticking point is that it will have to comply with the Australian spam 
legislation (which IMHO actually maps nicely onto IM2000 as it requires 
authorisation before any actual sending).

Lawrence