The Routing Over Large Clouds Mailing List Archive by date

Cell Relay Retreat>List Archive>month:1996-May> msg00193



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

TTL decremening (was Re: My personal take on cell switching routers)

  • From: kasten@ftp.com (Frank Kastenholz)
  • Date: Fri, 31 May 1996 08:49:34 -0400
  • Cc: malis@nexen.com, ion@nexen.com, gja@thumper.bellcore.com
  • Originating-Client: d-cell.ftp.com
  • Repository: mailserv-2high.ftp.com, [message accepted at Fri May 31 08:49:25 1996]

Tools like Traceroute and multicast-scoping implicitly rely on specifc
TTLs and actions to occur based on those TTLs.

TCP has some reliance on the upper-bound of the lifetime of packets.
Part of this is time-based (i.e. in 791 someplace it says decrement
1/second while in queues and 1/hop). If TTL isn't decremented
properly it may be remotely possible that a TCP connection could get
"weird". This is a very very low frequency event -- it might happen
once to any given person who would dismiss it as "a glitch" and retry
the operation and life gets back to normal...


BUT the real reason of my note is that this is all reminiscent of the
bridges (the things which forward packets at the level below IP) and
routers (the things which forward packets at the IP level) and
fragmenting bridges (aka extruders) and stuff like that. I guess we
have a new generation of networkers so we have to revisit all those
old discussions...


--

Frank Kastenholz