From: grahn+nntp@snipabacken.se   
      
   On Fri, 2020-03-20, groovee@cyberdude.com wrote:   
   ...   
   > I'm also curious about what exactly happens if a *faster moving   
   > packet*, say segment 2 of a stream arrives at the client in Mumbai   
   > before a slower, segment 1 (say via the Tajikistan link, while   
   > segment 2 came via Quicker Iran)?   
      
   Generally you try not to let related packets go different paths,   
   for reasons like this one.   
      
   > Will the Linux kernel just HOLD   
   > segment 2 forever inside RAM or something while waiting for the rest   
   > to turn up? (that's not a good thing, is it?) What about Windows?   
      
   You must be talking about TCP now.   
      
   I don't know, but it could either   
   - discard #2 and ack #1, forcing the sender to retransmit   
   - ack #0, but keep #2 around in case #1 shows up   
      
   I suspect the latter happens.   
      
   "Forever" sounded like a bad thing, but I don't think it is.  If the   
   local application is alive, if it hasn't requested any timeouts, if it   
   doesn't tell the socket to die, and if there's no explicit evidence   
   that the other side is dead, then the socket should be kept.  Even if   
   you're without network for a week.   
      
   /Jorgen   
      
   --   
     // Jorgen Grahn    O  o   .   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)