> karl - ICANN is gonna take on protocol regulation...isn't that pretty obvious?
No, ICANN is not going to be involved in protocol design or the
determination of what constitutes a "standard".
(I do expect, however, that eventually the DNSO will attempt to oversee
DNS protocols.)
It looks like from the latest noises that the IAB is going to makes its
own deal for that really tough job of writing down and publishing which
port number goes with which protocol.
So ICANN would not even be doing the job of writing down port numbers.
But even if ICANN were to undertake that job, most protocol parameter
assignements are "add one and write it down" -- raw clerical functions
devoid of public policy impact.
There are some number assignments which do take some technical knowledge
-- for example IP multicast addresses, MIME types -- but these are
increasingly being done by a new mechanism, "designated experts" to whom
the numbers assignment clerical function turns when questions arise.
(As an example of the degree to which hand waving is going on to obscure
the issue of protocol parameter assignment: One assignment is of values to
represent cryptographic hash functions. At first I was assured that the
IANA function would have to actually do a mathematical evaluation of the
strength of the hash. As it turned out, all that the clerical function
needs to do is ascertain whether the algorithm is published in a
cryptographic journal -- a job that can be done by simply asking a
librarian in the mathematics library of any university.)
The current ICANN thinking, as far as one can tell, is that ICANN would
simply deal with conflicts that might arise between registration of
protocol parameters by different organizations, one being the IETF.
Given that there have been relatively few, if any, fights between
standards bodies over which TCP port goes with which protocol, I feel safe
in thinking that this is a pretty empty playground.
And it is a playground not worthy of having three seats on ICANN's board
of directors.
--karl--
