[LINK] Fwd: UN hosts conference on safeguarding freedom of
expression on the internet
foconno1 at bigpond.net.au
Wed Feb 23 13:38:48 EST 2005
At 11:45 AM +1100 23/2/05, Markus Buchhorn wrote:
>Frank wrote (a long message, various points)
>>>which have contributed little to the cause of advancing interoperability and
>>>efficiency amongst systems (witness the X400 e-mail 'standard' or the X500
>>>directory 'standard' if you want concrete examples).
>I think there are two different discussions here - one is a
>conference on freedom of expression (Roger, you there?), hosted by
>the UN, and the subject of the original message. The other is the
>story that the UN is looking into a proposal for 'somebody' (else -
>ITU maybe) to take on ICANN's role(s). Different topics, and Frank
>is reacting to the latter.
Mmmm ... my bad. :)
>At 07:16 AM 23/02/2005 +1100, rchirgwin at ozemail.com.au wrote:
>>In the first half of the 80s, when ATM was first proposed, IP was a
>>research project of interest to a very small community. In 1989,
>>when the 53 byte cell was accepted, there was no consumer Internet
>>to speak of, just a larger research project.
>One story went that the frame-based network folks wanted larger
>cells (64bytes and up) to make all the fragmentation and re-assembly
>stuff easier and faster, and the phone folks wanted smaller cells
>(32 bytes) to handle echo-processing issues. In the best ways of a
>compromise, they averaged to 48 bytes, which made neither group
>happy :-) (ATM cells are 48 byte payloads, and 5 bytes of header =
>53). Another story had it split along the usual Atlantic divide,
Right. But the story is also that Cisco and the TCP/IP adherents
didn't get a look-in, and were atad miffed at the overheads such a
small packet size could cause ... especially with respect to a number
of TCP application protocols.
> >I do take your point, but it's also reasonable to observe that Big
>Iron similarly dominates IETF publications. Cisco, Microsoft, Nokia,
>Bell Labs - it's people like these that run the technical standards
>just as much on the net as in the ITU.
>The IETF working groups though (and most in the ITU) are very
>collegiate. The big-iron folks see issues more often than joe
>researcher, so they tend to put up the proposals first to solve a
>problem for them, and hence end up on the final documents. Many of
>the rfcs have big-iron and little-uni folks on them together, when
>the uni folks point out problems with the big-iron designs. Some of
>the big standards (like SIP, rtp, ...) come out of uni/research
>entirely. Sometimes they're updated by big-iron folks who found
>implementation problems, or want to add functionality. But anybody
>can equally submit IETF and ITU standards proposals.
Agreed ... so why do we need to restrict the process to the ITU
members. The current IETF (and I suppose ICANN ... which I still
dislike) protocols are available to one and all for comment and
>>Where the Internet has it all over the ITU is that its fast process
>>doesn't just let it introduce standards quickly - it can also bury
>>failures before they gather too much money and commitment.
>I think that's changed, in some areas. Phy-layer stuff is hard,
>slow, and outside the IETF scope. Things like protocols though can
>be just as quick through the ITU as the IETF (where i-d's can linger
>for years :-), and some rfcs never see serious implementation).
>H.350 (a schema for videoconferencing identity descriptions) went
>through the ITU in well under 12 months, and was almost entirely
>driven by university folks.
Mmm, the IETF basically takes responsibility for network layer (layer
3) and many transport layer (layer 2) protocols, and application
protocols (layer 1) basically go through a number of bodies (IETF,
W3C, OASIS etc). ICANN oversees Layer 2 and 3 protocol implementation
... but other than that doesn't have a huge role there.
More information about the Link