MAJOR ISSUE: Causes of "problems"

[Keith Moore]

> If a "problem" now exists where formerly there was none, a logical
> question to ask is "what changed?"
> 
> So I'm curious as to what people think some of the underlying *causes*
> of the "problems" are.

The Internet has gotten considerably larger and more diverse.  It is no
longer sufficient to design a protocol that will work in a benign and
fairly homogeneous network for a relatively small group of researchers.
The environment is no longer benign - serious large-scale distributed
attacks are a weekly occurance.  Network capabilities vary considerably
- available bandwith between two points can vary by six orders of
magnitude, maybe more.  Apps now have to content with NATs, firewalls,
interception proxies, and other obstacles that subtly disrupt their
communications and impose arbitrary constraints on when they can
communicate.  Users' interests have also become quite diverse - Internet
users no longer mostly speak the same language or use a uniform set of
applications (although almost everyone uses email and the web, the
uniformity ends there).  So now we see a demand to make things work over
IP that demand far more from the network than IP was ever designed to
guarantee; we also see conflicts between the different demands.

With the increased size we are also finding some limits to the
scalability of the original design - not only in address space size but
also in routing, and in (the lack of a) comprehensive security
architecture.  We have adopted piecemeal workarounds for some of these
problems, but these conflict with one another and with legitimate uses
of the network.

Also, to gain consensus on a protocol it is no longer sufficient merely
to find a way to solve the technical problems of how to represent
the data and exchange the information.  Vendors will now insist that the
protocol support or even mandate their business models - even when these
contradict each other or are not in the interests of users.  And lacking
the confidence to compete in the marketplace, some vendors have resorted
to using IETF as a battleground - blocking working group activity either
with political means or by the threat of IPR suits in order to coerce a
particular outcome or delay the result.

The low-hanging fruit has already been picked - having solved the easier
problems we are now tasked with solving much more difficult ones -
including problems for which there is no known solution.

Given all of the above, it's amazing we're able to do any useful work at
all any more.

What did I leave out?