[ih] Exterior Gateway Protocol

[Guy Almes]

Hi Noel,
   Your memories / comments are very interesting.
   Once again with a history (as opposed to technical) focus, I was 
particularly struck by two comments ...

On 9/3/20 6:14 PM, Noel Chiappa via Internet-history wrote:
>      > From: Guy Almes
> 
>      > While I'm not aware of anyone actually doing this, it is interesting to
>      > contemplate what an SPF-based *EGP*.  This could have advantages (e.g., more
>      > optimal inter-AS routing and rapid convergence times), but having any
>      > Global agreement on which inter-AS routes will be preferred would be
>      > very unlikely.
>      > So, in practice, SPF technology is only used in IGPs.
>      > ...
>      > Comments?
> 
> Tony already said something about this, but a few more notes.
> 
> 
> ...
> 
> An orthogonal subset of MD uses what we called Explicit Routing' (a name due
> to Yakov Rekhter, with thanks), where _individual_ nodes (perhaps recursively)
> select the entire path (or sections thereof, in the recursive case). This has
> several advatanges; in addition to being less fragile, and far less
> suscpetible to loops, it _allows_ (but does _not_ mandate) individal nodes
> to make their own decisions about paths - aka policy routing.
   Back in the 1980s, the Loose Source Routing and Strict Source Routing 
were typically observed by routers, so we had a sandbox with which to 
play with these ideas.  It seemed a reasonable and practical idea to 
make use of Loose Source Routing in applications where Policy Routing 
was demanded.
   Within a decade, of course, most any form of source routing was 
regarded with horror.


> 
> ...
> 
> Around then, I'd split off to work on Nimrod (the name is neither an acronym
> nor a backronym, but a private joke between me and John Lekashman, from whom I
> had hoped to get funding). The primary difference between IDPR and Nimrod was
> that I wanted to get rid of the EGP/IGP split - to me it was a blunt tool of
> limited capabilities. ...
   I take the EGP/IGP split to be, in essence, the basic Autonomous 
System (AS) concept.
   During the late-80s and early-90s period, the basic AS idea was 
amazingly powerful.  It allowed the wonderful variety of technical 
approaches to be taken by an explosively growing (by which I mean dozens 
or hundreds) number of campus, regional, international, etc. networks, 
while preserving a relatively simple inter-AS world where we had to 
cooperate.
   That allowed the Internet to grow into something amazing and 
unstoppable, if flawed.
   But you were right then and now in pointing out the weaknesses of 
this binary intra-AS / inter-AS split.
   I became particularly aware of that weakness when the IETF moved to 
expand the size of the AS field in BGP from 16 to 32 bits.  In 1987, I 
would not have anticipated 64,000 ASes.  Oh well.
   And then, again, when working with routers that had to usable in 
backbone networks and thus have a "full" routing table in expensive fast 
packet-forwarding cards.  If memory serves, an example was the need for 
10s of thousands of inter-AS routes in the forwarding card of the IBM 
routers used in the T3 version of the NSFnet backbone.

   Again, focusing on history, sometimes a technical idea (here, the 
binary intra-AS / inter-AS split, can allow an infrastructure to grow to 
become unstoppable and yet not be adequate for the additional growth it 
will need after becoming unstoppable.

   Cheers,
	-- Guy
> 
> ...
> 
> 	Noel
>