Hi Tony,

In "Re: [rrg] NANOG44 session on current IP address challenges", you
wrote, regarding map-encap:

> How does that reduce the table size?

Many or most end-user networks have micronets (roughly equivalent to
LISP EID prefixes).  These are not in the global BGP routing table.

Each Mapped Address Block (MAB) is advertised in the BGP routing
table.

Each MAB contains many (ideally thousands to millions) of micronets.

MABs are required so that OITRDs (Open ITRs in the DFZ - similar to
LISP Proxy Tunnel Routers) can collect packets addressed to these
micronets and then tunnel the packets to the correct ETR.  Without
this, there would be no support for multihoming and portability for
the adopters of the map-encap (really core-edge separation) scheme -
so hardly anyone would adopt it.

Because thousands (or perhaps millions) of end-user networks share a
single MAB, this greatly reduces the size of the global BGP routing
table compared to doing nothing, since without map-encap, many of
those end-user networks would gain their own conventional PI space,
and advertise it in the BGP system.

Map-encap is not the only approach - it is one way of getting
packets from ITRs to ETRs in a core-edge separation scheme.

Two other methods involve forwarding, using modified DFZ routers:

  ETR Address Forwarding (EAF) - for IPv4
  http://tools.ietf.org/html/draft-whittle-ivip4-etr-addr-forw

  Prefix Label Forwarding (PLF) - for IPv6
  http://www.firstpr.com.au/ip/ivip/ivip6/

Unlike map-encap, these have no encapsulation or overhead or PMTUD
problems.  Without the PMTUD problems, the ITRs and ETRs can be simpler.

Also, there is translation - Six/One Router, for IPv6 only.

  - Robin

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

Reply via email to