RE: Single vs many solution(s)

Yakov,

[SNIPPED]

>> The one that reflects reality, and not the one that reflects wishful
thinking. With this in mind the following from Ross' e-mail answers
your question. <<

As I have stated before I am generally sympathetic to the idea that there
are very different service providers with very different requirements; and
therefore perhaps multiple standards are called for. But, some comments.

I did think Ross's email was excellent, but some further color to his
comments might be beneficial.

In the case of developing RIP, OSPF, BGP4, etc. we kind of lucked out (or
perhaps it was good intentions) in the sense that all three could be used
easily in the same network - and together - because it was relatively simple
to export/import routes (where applicable and advisable) because they all
used the same addressing structure, so a route was a route was route which
could be imported into the kernel (or its equivalent). So things worked out
nicely. But what if they had all used different addressing structures, or
where othewise incompatible?

In the case of the Ethernet example, things actually did not work out so
well. First it has to be noted that even the IEEE recognized the need for
some level of interoperability, so IEEE 802.2 was defined for all
(Standardized by the IEEE) Ethernet MACs. But even this did not produce good
results. Bridging between Ethernet, TR, and FDDI was **very** painful. There
are many reasons, some related to some interesting layer 3 encapsulations,
in the case of IPX, the old canonical/non-canonical issue, function points,
blah, blah, blah.....etc. Now we can argue that eventually it all worked out
well because the industry trended in one direction, but for the peroid where
vendors had to support all three, it was extremely painful; on both an
interoperability and chip level (in the case of TR); and all three had to be
supported because their vendors could claim the legitimcay of being
standards.

So at a minimum, it would be nice to see obvious (meaning it may be occuring
but not visible) evidence of cooperation between competing proposals to
identify areas of commonality, and seek interoperability. More generally,
such a process may lead to other levels of cooperation.

To date we have asked the binary question of single or many. Is there also
some intermediate states?

Best,
Mark