I have updated the Summary and Analysis document:

  http://www.firstpr.com.au/ip/ivip/Ivip-summary.pdf

to cover the two alternatives to encapsulation:

   ETR Address Forwarding (EAF) - for IPv4
   Prefix Label Forwarding (PLF) - for IPv6

The rest of the document has only minor improvements.  The new
material is at the start and the end.  Older versions are still
available - there is a changelog at the end.

Ivip began as a map-encap proposal.  However, it is no longer correct
to describe Ivip as being a map-encap system, or to regard it as
being closely related to LISP, APT or TRRP.  There are some
similarities with APT and LISP, but the differences are probably more
important than the similarities.

The long-term endpoint of Ivip, for both IPv4 and IPv6, is to use the
above "Forwarding" arrangements, rather then encapsulation.  This
removes the need for solving some nasty PMTUD problems which are
inherent in encapsulation.  Consequently, ITRs and ETRs can be much
simpler than with encapsulation.  Forwarding is 100% efficient, since
there is no encapsulation overhead.

Ideally, Ivip would be implemented purely with "Forwarding".  This
requires modest upgrades to essentially all DFZ routers and some or
many internal routers.  Any router which classifies and forwards its
packets with firmware, rather than hardware, can surely be upgraded
to perform the new functions.  These changes do not affect BGP at all.

Ivip is best thought of as a core-edge separation solution to the
routing scaling problem, with an optional map-encap mode if it is to
be deployed before all DFZ routers can be upgraded.

If it is deployed with encapsulation, all ITRs and ETRs will be
capable of using "Forwarding" as soon as the DFZ routers are upgraded.



 - Robin
_______________________________________________
rrg mailing list
[email protected]
https://www.irtf.org/mailman/listinfo/rrg

Reply via email to