ietf
[Top] [All Lists]

Re: Internet 2020 Goals

2014-05-27 13:02:47


On 5/26/2014 7:15 PM, Miles Fidelman wrote:
At one time, NSF and the National Academy of Sciences used to publish
"research agendas" for networking - when did this stop?

It didn't, and all such reports (including 2020 visions) have had the same amount of impact (lots of heat, but no light IMO).

Unless the IETF is prepared to put up research money to back it up*, making statements about the future of the Internet is unproductive prognostication.

Joe

Carlos M. Martinez wrote:
I agree with almost everything you wrote, and I think we definitely need
something akin to a corporate 'vision' but for the Internet.

What is our utopia? What do we want the Internet to look like and how do
we want it to work in 5, 10, 15 and 20 years ?

I also agree that some of these goals might not be IETF-related, but I
think we (as the IETF we) can set a vision of what we would like the
Internet to be.

cheers!

~Carlos

On 5/15/14, 1:57 PM, Phillip Hallam-Baker wrote:
While we frequently see complaints about where the Internet is headed
on this list and sometimes see proposals to change that direction we
seem to rarely consider what that direction should be.

The only time we have done that recently at a high level has been in
response to the Snowden disclosures. And we will probably revisit that
in the near future when Greenwald drops the final installment of the
documents which is who is being targeted.

Privacy is one important goal that we want to achieve. But it isn't
the only goal that I think is important. And I don't think we will
make much progress even on Privacy unless there is some sort of
deadline.

Hence the need for Internet 2020. We have six years to go before 2020
which is sufficient time to do useful development and deployment. It
might even be possible to do research, development and deployment in
that time frame.

Right now we have a lot of governments that are trying to get involved
in the governance of the Internet. Often without any understanding of
what it is or why it is built the way it is. A risk we face is that if
we do not define what our goals are, others will do that for us.

Another risk we face, one that I think is unfortunately realized
rather often here is that if we don't state our goals, we can't make
common cause with others who have the same set of goals. Most IETF
participants would probably agree with most of what emerged as
consensus at Netmundial. But most IETF-ers still regard the initiative
as a threat rather than an opportunity.

If we set concrete goals and realistic deadlines then there is a
probability that at least some policy makers and providers of research
funds will factor those goals and deadlines into their own decision
making processes. This is how the policymakers themselves work. The EU
did this when they set 1992 as the date for 'completion' of the single
market. Now as everyone who is familiar with the issue knows, 1992 was
neither the start or the end of that process. But setting that date
did act as a forcing function that drove the process from 20% to 90%
complete.


So what are the goals that we might reasonably hope to achieve by 2020?

By goals I mean providing capabilities that meet user requirements,
not the deployment of specific technologies. So I don't see deployment
of IPv6 as the type of high level goal we should start with.
Deployment IPv6 is not an end in itself, deployment of IPv6 is merely
a means to the end goal of ensuring that everyone can get Internet
service. And with a population of 7 billion we obviously need more
than 4 billion IP addresses.

So IPv6 is really just one technology that is serving the high level
goal of Access: The Internet is for everyone. But achieving the access
goal means much more than deployment of IPv6.

Returning to the policy maker interface. One of the problems with IPv6
deployment today is that the deadlines have always been fungible. Even
now with the IPv4 address space exhausted, IPv6 is far from
ubiquitous.


The three high level goals I see are:

1) Security: All Internet protocols should provide confidentiality and
integrity by default.

2) Access: The Internet is for everyone and everyone should be able to
use it regardless of their geographic location or political
interference.

3) Autonomy: [Here I need a concise definition]


The third goal is the hardest to define but probably the most
important. One of the main reasons that the Web is designed the way it
is and why it became so popular was that it gave individual users
control.

At the time the Web was invented, the IT infrastructure at CERN was
centered on a large mainframe system, CERN-VM. The mainframe
administrators were the aristocracy and the users were serfs. A lot of
time and effort went into reminding the users of this. Anyone who got
an account on CERN-VM would have all their mail sent to that machine
by default, that was not a choice. And since it was an ESIDIC machine
it meant that your mail would never work again.

The Web broke up the monopoly of access to the information hoarded on
CERN-VM. It wasn't necessary to have an account on the mainframe to
read the phone book or to find out the Delphi meeting schedules or any
of the other inconveniences that the commissars of CERN-VM had
inflicted.

The Web was designed to break up that information monopoly which is
why it then went on to break up the planned information monopoly of
'Interactive-TV', the future of networking as imagined by Time Warner.
The only interaction in Interactive-TV was that you could buy stuff
from overpriced online shopping malls. Users were just passive
consumers. The designers never thought about about the possibility
that users might actually interact by creating content themselves.

But now as a result of institutional design and network effects we
face forces that threaten to eliminate user autonomy returning us to
the pre-Web model.

My problem with the institutional arrangements in ICANN is that I have
no recourse whatsoever. If ICANN decides to jack up the price of
domain names to $100 each, I have to pay up or lose my domains. And
that is important because when I am using hallam(_at_)gmail(_dot_)com, Google
effectively owns my Internet email presence. If they decide to change
terms of service or impose ridiculous fees then exercising my exit
right of switching email providers comes with an enormous switching
cost.

So the starting point for Internet autonomy is to have our own domain
names. Which is why I own hallambaker.com. But while this gives me
autonomy with respect to google, I do not have autonomy with respect
to VeriSign or ICANN.

DNS raises important autonomy issues which is why I have been working
on it for the past five years. But it isn't the only area where
autonomy is an issue.

Autonomy is an important framing on the net neutrality debate. I find
the usual framing highly suspect because there are circumstances where
I would be happy to have free Internet access that doesn't work with
Vonage or Netflix unless those companies pay the provider. My real
concern is whether that is a choice or a mechanism to extract unfair
rents where I pay the ISP for the Internet service and then pay again.

Autonomy is also the reason that I am very concerned by the recent
trend where web properties require the use of one particular social
network to comment on their blogs. The commercial reason they are
doing it is of course that they want the exposure that comes from
linking the comments to that social network and advertising the
comments to their friends. But they are taking away my autonomy and
they don't even register that some people would legitimately see that
as unacceptable.

And lest people accuse me of being anti-corporate here. Time-Warner
made far more money from its Web properties than it would ever have
made from their Interactive-TV. And the Time-Warner management that
was clueless about Interactive TV also managed to burn half their
shareholder value by gifting it to the shareholders of AOL, a business
unit they eventually disposed of for a few percent of what they paid.

My thesis is that companies that back the goals Internet users want
will be the leaders of the Internet in 2020. So if I am right and
Security, Access and Autonomy are the goals they will get behind,
companies that want to become or remain leaders should back them.


So those are my goals. What goals should we be attempting to address?
What are realistic timescales?

Are we just going to be happy with a faster Internet with an
effectively unlimited address space or do we have bigger goals?





<Prev in Thread] Current Thread [Next in Thread>