From: grahn+nntp@snipabacken.se   
      
   On Tue, 2010-09-28, Andy Furniss wrote:   
   > Jorgen Grahn wrote:   
      
   [about some third guy tuning MSS down]   
      
   >> You've mentioned this several times recently.  What about PMTU   
   >> discovery -- doesn't that work for you?  It seems to work well enough   
   >> (or never kick in) for me.   
   >   
   > Does your ISP restrict your inbound to < 1500, though. If it doesn't you   
   > are not likely to have problems.   
      
   Thus "never kick in", perhaps. No, my link is 1500. But I'd expect to   
   invoke PMTUD now and then when I'm talking to some other host with a   
   more narrow link. Maybe not when visiting major web sites, but when   
   file sharing and so on. (On the other hand, there are always   
   Bittorrent peers which have problems ...)   
      
   > In the case of pppoe type ISPs the 1492 they (may) run is overcome by   
   > the fact that consumer dsl modem/routers mss clamp TCP avoids PMTUD   
   > black hole faliures.   
      
   Ugh. I've heard about MSS clamping, but I thought it was rare.   
   As I understand it, it means   
   - they have to modify your IP payload   
   - they have to keep per-connection TCP state in the "router"   
   - it's invisible to debugging, unless you have control over   
     both endpoints and compare what A sends to what B actually   
     sees.   
      
   Does it really benefit anyone?  Wouldn't an ISP be better off not to   
   break PMTU discovery in the first place?   
      
   /Jorgen   
      
   --   
     // Jorgen Grahn    O  o   .   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)