3b7c3e9e   
   From: barmar@alum.mit.edu   
      
   In article   
   ,   
    David Schwartz  wrote:   
      
   > Thank about it logically -- it is impossible any hop to show worse   
   > than any hop after it. So if you think hop 3 is somehow worse than hop   
   > 4, it is an artifact of how you are measuring. (Because you have to go   
   > through hop 3 both ways to get to hop 4.)   
      
   That's only true if routing is symmetric.  If the path back from hop 3   
   goes through a congested link, but the router at hop 4 knows an   
   alternate path that bypasses it, you can get better latency from a later   
   hop.  Asymmetric routing is extremely common, since most routing   
   protocols don't make any attempt to prevent it.  This is especially true   
   if the path uses multiple autonomous systems.   
      
   This is one of the main limitations in using traceroute, it only shows   
   you the path in one direction.  To be sure of what's going on, you need   
   to run traceroutes from both ends.  Back when I worked for an ISP, we   
   used the -g option to traceroute to force it to go through a particular   
   remote router and back to us, since it was usually not possible for us   
   to login to something at the other end.  But normal ISP customers   
   generally can't do this, because customer access routers usually block   
   packets with source routing options.   
      
   Also, routers are generally much faster at forwarding packets than at   
   sending ICMP Time Exceeded messages, so in some cases the extra delay at   
   a hop can be due to the router's main CPU being overloaded, not anything   
   having to do with the links.   
      
   --   
   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)