Harald.
Interesting, important, thanks.....
Internet usage
==========
One of the large dynamics not explicitly mentioned is the increased
commercial usage/value of the Internet and how that drives the community
in new directions.
Trust model
=========
Inherent in Eric's problem statement is the notion that end systems have
the ability to impact the experience other Internet users have. Whether
this is the result of an historical trust model, where people using the
Internet were assumed to a) have clue and b) be acting in the best
interests of the community, or whether this is the result of other
community values, this diserves comment/debate, IMHO.
The basic dynamic of course that needs to be balanced is the rights of
Internet users, the role and obligations of operators, and the creative
abuse of the network that could lead to unforseen new value and
applications.
Discusion of trust models will inevitably also lead us down the road of
discussing hairy issues such as (D)DOS - but isn't that one of the more
pressing issues today?
Network Architecture & layers 1-9
=========================
Firstly, seems there are four basic functions that need addressing in
the building of internets (remembering the network of networks value,
there is not one "I"nternet):
-Transport of packets from one edge of an internet to the other edge
(PE if u like, but...).
-Transport of packets from end system to end system over one of more
internets
-Management and operation of internets
-Applications that make use of the above infrastructure
Based on the inclusivity problem statement positions you have
considered, and other comments WRT how much the IETF can do, it would
appear that over time the last bullet point above might become disjoint
from the first three. Not arguing that application development should be
done at the IETF, but I think that some recognition needs to be made of
the vastly different skill sets and interests of end-system operators
and engineers and infrastructure operators and engineers. Perhaps the
"area" concept needs another super-layer, especially given the vast
difference in communities of liasion.
Lastly, some discussion of layer 1-to-9 should take place. Seems to me
the IETF works very well when focused at layers 3 to 4; has of course
established many important application layer standards; but experiences
challenging liasion when focusing below layer three (data/user plane) -
PPP not withstanding. In the spirit of "we can not do everything", this
is deserving of discussion if for no other reason than to clear the air
as we move forward.
Perhaps one of the most important areas of focus is the connection of
end systems and internets to other internets. This is an area that must
allow for creative abuse of the network through both freedom of higher
layer protocols and also the facilitation and participation of as many
systems vendors as possible addressing cable, wireless, wireline,.......
This area should especially be supported by standards, implementation
agreeements, and interoperability efforts.
The immutable vs the mutable
=====================
Seems like it would be useful to separate the immutable from the mutable.
Examples of immutable might be (not a prescription just examples):
-rough consensus
-running code
-Chairs that operate in the best interests of the community
-the UNIX-like adherence to development of small building blocks (also
present in biology BTW).
-network of networks
Examples of the mutable might be:
-over the next 5 years we are going to focus on (D)DOS
Lancing the boil
============
There is nothing nice, pleasant, or enjoyable about lancing boils, but
if we are to do so, then the starting point should be the encouragement
of the identification of the largest boils.