--On Wednesday, March 20, 2013 23:36 +0100 Jari Arkko
<jari.arkko@xxxxxxxxx> wrote:
> I think it is mostly market forces and historical reasons, and
> the development of the IETF to focus on more particular core
> aspects of the Internet (like routing) as opposed to what the
> small shops might work on.
I mostly agree. However, I see lots of activity in Apps and
RAI, very little of which would seem to be "core aspects of the
Internet". Also, given the cost factor, the length of time it
usually seems to take us to spin up a WG and get anything done
is probably also a significant barrier: a small shop who could
afford to send someone to a meeting or three might have neither
the people-resources nor travel and meeting budget to commit to
a few years of meetings.
> But I think we are missing a bit of the point in this
> discussion. I do not feel that we need to prove we are somehow
> "no worse" than industry average. The point is that *if* we
> had more diversity along many of the discussed lines, we'd be
> far better off. For instance, having people from multiple
> organisations provide input to a last would be preferable to
> just a few. Similarly with the other dimensions of diversity.
> When I talked to some of the ISOC fellows last week, I
> realised peering is very different on different continents.
I have run across another example fairly often that was, I
think, mentioned briefly last week. Most of us are used to
network connections of very high bandwidth and quality. Our
protocol designs and implementations are developed and tested on
those networks and, usually, on the very latest and most
powerful equipment. The IETF would almost certainly benefit
from vigorous input from people whose environments are
characterized by longer delays, serious congestion, packet
fragmentation, and so on. Without them, I fear that
implementations of a lot of our work, and maybe the work itself,
will not be acceptable in lower-end or more congested networks
or from computer systems more typical of that average Internet
user and that they will not have the level of robustness that
ought to be one of the Internet's strengths.
> Even if there may be less economic activity on networking on
> those continents, it would be good for us to understand the
> real situations around the world, as opposed to thinking the
> whole world is like where we live. Diversity = good in most
> cases, and increasing that goodness should be the goal.
Indeed. And, taking the comment above a step further, one need
only go to a sufficiently rural area in many developed countries
--one that is served exclusively by overloaded high orbit
satellite links (with the delay times that implies) -- to
encounter the problem even though people from "other continents"
are more likely to be articulate about the issues and easier to
get to the IETF then those rural users.
john