Henning Schulzrinne wrote:
I think this whole discussion would benefit from some concrete examples.
What wholly new protocols has the IETF developed in the past decade?
Which ones would you consider successful or not?
Well, that's such a reasonable question, I did a subjective review of Proposed
Standard RFCs for the last number of years -- ignoring that most recent and
going back to rougly RFC 2500 -- looking for acronyms that were for
significant IETF-generated efforts.
Note that I said subjective. I meant it two ways. One is that I merely did a
scan, rather than anything more diligent. The second is that the list
reflects my personal reactions as to 'significance' of the effort:
IPv6
DNSSec
Enum
XMPP
CPIM
BEEP
MSRP
SCTP
IPSec
SASL
MPLS
CalDav
RTP
NETCONF
IMAP4
PPP
(a variety of PKI efforts?)
LDAP
OPES
VPIM
FFPIM
GEOPRIV
Diameter
iSCSI
DDDS
SIEVE
A number of questions come to mind:
1. What additions or removals should be made to the list?
2. Has the protocol (or set of protocols that pertain to the acronym) gained
widespread adoption and if so, what does "widespread" mean and what does
"adoption" mean? Having 100M+ daily users is definitive, but what about
smaller numbers that nonetheless dominate an important niche, such as
"LAN-email office workers"?
3. ...?
d/
Almost by necessity, newer protocols tend to cover niches, relatively
speaking, as opposed to broad swaths of technical territory
("asynchronous message-based communication", "object retrieval", "name
mapping").
Henning
On Dec 31, 2007, at 3:09 PM, Dave Crocker wrote:
Iljitsch van Beijnum wrote:
On 28 dec 2007, at 7:41, Franck Martin wrote:
The "What makes a protocol successful" presentation, shows that the
best protocols are the ones given to IETF for it to refine and
complete. They have already a user pull when they reach IETF.
I don't think that's valid statistics: obviously many of the
protocols in question were already successful before they were given
to the IETF, which isn't necessarily the case for protocols developed
"in-house".
That's the point: protocols created in the IETF, over the last 10
years, do not have a very good record of deployment and use.
Protocols created outside the IETF and then brought into the IETF have
a better track record.
This suggests that the IETF has become a place to refine existing
(proven) work and to expand its community of adopters, rather than to
formulate initial work and gain initial adoption.
On reflection, this should not be viewed as heretical or even
problematic.
When a group grows to have the kind of massively diverse participation
that the IETF now gets, it is rarely going to be possible to develop
or maintain the kind of coherent focus needed to get solid, cohesive
specifications written: Things move from getting "different
perspectives" to getting "constant objections".
A successful protocol needs a tight design focus and a a clear market
need. Our current base of participation makes it difficult to
formulate either.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf