t-delegated.example.com> 25fcb672   
   From: grahn+nntp@snipabacken.se   
      
   On Fri, 2010-09-10, Barry Margolin wrote:   
   > In article ,   
   >  Jorgen Grahn  wrote:   
   >   
   >> On Thu, 2010-09-09, glen herrmannsfeldt wrote:   
   >> > In comp.dcom.lans.ethernet Jorgen Grahn  wrote:   
   >> > (snip)   
   >> >   
   >> >> UDP-based protocols keep state too, and are just as picky about where   
   >> >> the datagram came from as TCP. You cannot expect every datagram to   
   >> >> contain full context, so the application protocol uses the source   
   >> >> address:port (and maybe destination too) as a key to lookup state for   
   >> >> the "conversation".   
   >> >   
   >> > Some UDP protocols are picky, but most aren't.  TCP identifies   
   >> > a connection by the quad   
   >> > source-address:source-port:destination-address:destination-port.   
   >> >   
   >> > Many UDP protcols/implementations will accept anything coming   
   >> > into the appropriate port.   
   >>   
   >> I can imagine that applying to DNS and maybe old-fashioned NFS, but do   
   >> any /relevant/ protocols do that?  Let's say those that aren't (cannot   
   >> be) stateless?   
   >   
   > UDP is rarely used for stateful protocols, it's mostly for idempotent   
   > request/response protocols (e.g. DNS).  The only counterexample I can   
   > think of off the top of my head is TFTP, and it's rarely used across the   
   > Internet.   
   >   
   > What "relevant" protocols are you thinking of?   
      
   That's part of my problem -- which UDP-based protocols are in use   
   today and "need" the anynomity, and would they break by the Skybuck   
   change?   
      
   I can enumerate the UDP-based protocols I know:   
   - DNS   
     would work   
   - NFS   
   - TFTP   
     aren't used on the open internet   
   - L2TP   
   - GTP versions 0--2   
     tunnelling protocols, stateful, would break, don't need the change,   
     aren't used on the open internet   
   - RADIUS   
     doesn't need the change,   
     isn't used on the open internet   
   - RTSP   
     streaming sound and video, would break I think, doesn't   
     need the change   
   - BitTorrent over UDP   
     could use the change, but I assume it would break (surely it   
     cannot be stateless?)   
      
   (Of course, the idea is broken for other reasons too, so in a sense   
   I'm flogging a dead horse here.)   
      
   It seems to me that you can discuss TCP in a general sense and get   
   something meaningful out of it, but you cannot discuss UDP in general   
   -- you have to talk about a specific use of it, because most of what's   
   interesting happens one protocol layer up.   
      
   /Jorgen   
      
   --   
     // Jorgen Grahn    O  o   .   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)