[ih] BGP vs EGP

[Noel Chiappa]

    > From: Bob Braden <braden at ISI.EDU>

    > Early efforts by Mills and others to extend EGP into an inter-domain
    > routing protocol failed.

I recall this vaguely - it was not just Dave, IIRC, I think there was a whole
group set up - I was there, I seem to recall Martha, and I'm pretty sure I
remember Mike St John, but there were others too.

I don't think there's anything online about it, but I probably have some
hardcopy upstairs somewhere if anyone really cares. It was called EGP3, if
memory serves.

I don't recall exactly why it never went anywhere - I think it was a combo of
it being pretty complicated (a touch of 'second system' disease), of some of
us going off of the whole destination vector approach anyway, but maybe there
were other factors too.


    > There was an effort to design a global link-state routing protocol

Two, actually (or maybe three or four, depending on how you count).

There was an abortive effort by a group centered around Proteon to define a
simple link-state protocol to replace EGP - that effort was nixed by (IIRC)
Dave Clark because he thought it would be a diversion for Proteon and/or
outside their skill set. That effort was just as the EGP3 thing was clearly
grinding to a halt, IIRC. I have this vague memory that it was called FGP (in
the tradition of B, C, etc).

Then there was IDPR, based on the output of the Open Routing Working Group
(see RFC-1126), which was sort of a pre-Nimrod (it had a lot in common with
Nimrod, but was an EGP only - I wanted a routing system that ran routing from
the hardware, all the way up).

Then there was Unified, by Yakov and Deborah Estrin, which tried to give the
best features of both link-state and destination-vector by a combo of both.
And finally there was Nimrod.

(Note that IDPR != IDRP - the later was a BGP variant/proposed-successor that
handled both IPv4 and NSAPs, IIRC. A lot of its ideas wound up in later
versions of BGP, IIRC.)


    > but Eric Rosen (BBN) turned out a hard-to-read but ultimately
    > convincing note about why this would not work.

I don't recall that? When was this, if you recall?

	Noel