Re: [Privsec-discuss] [Stackevo] Consolidation

Hi, Jari,

One point, that you didn't ask for feedback about ...

On Sep 29, 2017 05:00, "Jari Arkko" <jari.arkko@piuha.net> wrote:

In the summer, the IAB briefly discussed the consolidation trends around
the Internet. Consolidation may take many different forms, for instance in
terms of traffic flows becoming more focused on large content providers;
consolidation in the ISP industry; drive towards fewer but more popular
operating systems or platforms; consolidation in the DNS or CDN industry;
dominance of popular applications and types of accounts users have; limited
sources for applications; and so on.

We all probably recognise at least some forms of these trends. In general,
some of these changes are a part of the Internet becoming globally
commercial, the best solutions winning market share. But there are of
course some concerns as well, starting from the effect few monocultures may
have on security issues. Or to ensure openness, and the ability to ensure
that innovation continues to generate new, better approaches in the
Internet.

But it is worth distinguishing architecural issues from marketplace
problems. Obviously, much of this space is outside the scope of the IAB and
more in the area of competition policies and economics. But there may be
questions for the IAB as well, whether there’s something that we should
understand better, or perhaps document.

I would have agreed with what you said here before the IRTF added GAIA and
HRPC, but given that those topics are now in scope for the IRTF that you
folks on the IAB have oversight responsibility for, "competition policies
and economics" doesn't seem as far out of scope to me now as it would have
three years ago.

That observation may not make any difference for your current investigation
on the three questions you asked about, and could be either wrong or
premature, but I thought it was reasonable to say that now.

For instance, many of the moves towards consolidation are driven by
economic factors rather than technical factors, but that it may be worth
studying to see if there are any assumptions about the architecture of the
Internet that no longer hold true. Similarly, there might be research
questions about the state of the Internet that the research community
should pay more attention to. For instance, research on Internet traffic
flows and how those change wrt their diversity and/or concentration over
time. As an example, is there a more recent version of what we saw in 2010
at the IETF 77 plenary: https://www.ietf.org/proceedings/77/slides/
plenaryt-4.pdf?

So, do the folks in the different programs have advice to the IAB in this
topic? In particular:
1. Do you see architectural issues within your own field that relates to
consolidation trends, and how those affect the Internet?

It might be worth noting that our increased reliance on user-space
transport protocols may be taking us in the opposite direction from the
trends toward consolidation elsewhere for the internet.

One doesn't have to control kernel TCP and SCRAP implementations for hosts
with a large deployment footprint to make the Internet behave differently.
See QUIC for a soon-to-be worked example (not the current chartered NewReno
direction, but the first alternative congestion control scheme, which won't
surprise me at all when it comes out) :-)

2. What potential new architectural structures might provide support to
either fighting the consolidation trend, or accelerating it?

3. Are there research programs that you believe would be useful in this
space, but are currently not being pursued? Which ones?

Perhaps, but see my note about "competition policies and economics" above.

Spencer

Jari Arkko for the IAB

_______________________________________________
Stackevo mailing list
Stackevo@iab.org
https://www.iab.org/mailman/listinfo/stackevo