6fea6ac0   
   From: barmar@alum.mit.edu   
      
   In article   
   <8102145a-e960-4ead-bbd9-8e7430870696@z35g2000prh.googlegroups.com>,   
    David Schwartz  wrote:   
      
   > On Dec 17, 10:50 am, karthikbalaguru    
   > wrote:   
   >   
   > > Interesting :-)   
   > > Can't there be some factor like channel condition / session condition   
   > > (These values determine the reliablitiy) ? I think, this factor can   
   > > be considered to make it dynamic.   
   > > Is it feasible ? Any thoughts ?   
   >   
   > You're not making any sense.   
   >   
   > It *IS* dynamic. First it's X, then it's 2*X, then 4*X. That's   
   > dynamic.   
   >   
   > Whatever algorithm we choose, once we choose that algorithm, that's   
   > what it is.   
   >   
   > DS   
      
   And the "session condition" is "there's a problem that caused a   
   timeout".  It isn't really feasible to determine a detailed reason, the   
   Internet simply doesn't work that way.  The assumption is that it's   
   usually due to some node being overloaded, so anything that required   
   sending explanatory data back to the sender would make things worse   
   (this is why ICMP Source Quench is not recommended).  So all you can do   
   is infer the problem from the timeout, and then slow down to reduce the   
   overload.   
      
   --   
   Barry Margolin, barmar@alum.mit.edu   
   Arlington, MA   
   *** PLEASE don't copy me on replies, I'll read them in the group ***   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)