Joho the Blog[berkman] David Clark - Should the Internet have a future? - Joho the Blog

[berkman] David Clark – Should the Internet have a future?

David Clark, one of the inventors of the Internet (he says he’s not a “father” of the Internet but is maybe a first cousin), is giving a lunchtime talk at the Berkman center.

He says that the decisions that will shape the Net are not made by the techies but by the world where business and economic interests reign. Already many of his colleagues are stepping away from the Internet “so they can have some space to innovate.” E.g., a friend is putting sensors in the forest instead of in cities so she won’t have to face privacy issues. [Yeah, but once squirrels get lawyers, she’ll be sorry. Plus it’ll all have been recorded by the sensors. Slam dunk.]

The research community should stand up and announce objectives for the Internet. A solicitation is about to come out from the National Science Foundation for “Future Internet Design.” It’s a challenge for the research community to come together about what the Net in 10-15 years should be, and then propose the research required to get there.

We need to do this in order “to come up with an architecture that has a coherent framework for discussing security.”

Isenberg suggests that this is an end-to-end issue, not something we want to build into the Internet.
DC: “Does anyone other than a geek” believe that security is a matter only for the application layer? Congress doesn’t believe that, DC says. We’re being “simplistic and unresponsive” if we keep saying that security is someone else’s problem. If we don’t do anything, Congress will pass a law, probably putting burdens on the ISPs. The ISPs will be given the job of policing your machine. No encrypting will be allowed, for example.

DC’s main point is that we need to be thinking about the architecture of the Internet, not thinking about incremental mods and bandaids. The forces that bring real change will not be technical. They will be the social concerns, policies, economics, business, competition…All these forces need to be at the table.

Zittrain: If the line between the PC and the network is so thin, is the NSF challenge really to think about the architecture of the PC as well of the Internet? And given the constellation of players in the PC field, is that practical?

DC: To be practical, I take some things as invariant: E.g., we will never have bug-free machines/sw.

Me: But doesn’t putting businesses at the table mean that, for example, Microsoft will insist that – as the majority supplier – the Net ought to work in ways that work with its software? And that may not be in the best interests of the Net.

Zittrain: And this runs against the old way that says that if the Net is open to any app, we can work these things out [rough paraphrase]

Charlie Nesson: I’m surprised, DC, to hear you say this because you were the person who rejected kings in favor of “rough consensus.”

DC: I got an empassioned email saying that what I’m proposing is incredibly stupid because any degree of centralization will play into the hands of the Evil Empire. But I think it’s worth asking what we want the Net to look like in 10-15 years.

Rebecca MacKinnon: Isn’t this too US-centric?

DC: NSF knows it’s US centric, so they’re trying to get some funding from other countries. But most of the research does come from “first world” countries. And you personally shouldn’t perceive this as an outsider talking to someone else who’s going to make a decision; you should be at the table. (But first, he says, you have to learn to talk tech talk, e.g., know what a port is.)

DC: We should fight to preserve the rights of individuals on the Net. Right now, it’s tipping against that. E.g., CALEA. We will not be designing the future but rather designing the playing field.

When the Morris worm came out, DC got a call from a DARPA colonel who asked “What should I tell my superior officer?” DC said: That the Internet fulfilled its design spec by delivering the virus to all machines at maximum speed.

Isenberg: The real value of the Internet is its “option value”: The ability to innovate. The people at the table, though, don’t care about option value. They care about how to catch the bad guys, how to make money selling stuff, etc. I’m afraid that the red machine [the virtual machine that is able to roam the Net freely, as opposed to the green one that runs authenticated, safe apps – a proposal gaining currency] will end up being used only by the hackers, so the cool new options aren’t available…

DC: I’ve talked with consumers and they don’t have any idea what option value is about. They say they’re terrified some computer zombie will delete my photos. They don’t care about the open future when compared with the current value. That will lead to a tremendous force to constrain the Net. We need a social analysis of this system, not a technical one.

Amanda Michel: Your timeline is 15 yrs but there’s pending legislation. How do these timelines meet?

DC: I don’t know.

Simpson Garfinkel: We haven’t been able to get IPv6. How are we going to get the sort of change you’re looking for?

DC: All of IPv6’s features except for the 64bit address space have been retrofitted into IP4. So, IPv6 has been a success. For security, that’s harder because you probably can’t get there piecemeal.

Me: Why do we want people at the table whosse values are not the values of the open Internet? [condensed]

There’s a way of thinking that says we only make progress by bringing together opposites, and that doesn’t allow coherent thinking about design.

Me: To me, it’s like you’re asking a comittee to come up with a compromise on what I think is a basic right. So, when you say we should bring to “the table”, what table is it? Is it merely a metaphor?

NSF is soliciting research from academics. Others should be involved.

Isenberg: I’m afraid that the big “stakeholders” will come but the disadvantaged won’t.

Simpson: So you [not DC] are proposing not using democratic processes, because if you voted by nations, the future design would not be open.

DC: This group of researchers that might be convened by the NSF has absolutely no power. The interesting question is why would anyone pick this idea up? The research community might be able to push with a better of idea who they’re doing this for.

[Note: As DC was leaving, we chatted briefly, attempting to “debug” the conversation, as DC says. The problem seems to be this: Becauses DC introduced this by talking about bringing economic interests, etc., into this, some of us (= me and others) assumed the brunt of DC’s suggestion was that we bring Cisco, Microsoft, etc. into it. In fact, his aim is to bring social activistis into it so that the next gen of the Net isn’t designed purely by American techies.]

One Response to “[berkman] David Clark – Should the Internet have a future?”

  1. David: this is a very interesting forum. I am surpraised that there we no IETF/IRTF fellow involved. Albeit, I have to take it , that it was an ‘academic’ discussion.

    Technically speaking the nextgen Inet was a consortium of universities and research org’s that worked with NSF w/VBN (veryhigh speed backbone network.

    IPv6 tunneled over ATM and/or nat’ed V4 addreses are a nigthmare for security. While the “rest” of world is moving into this paradigm, leaves the american techies wondering how to bridge the gap for the interim measure. Afterall, by 2015 most deployment would be v6 net.

    So the core question here is how the heck is the americas going to protect and monitor for the sake of security this type of meshed cloud ?

    So “his aim is to bring social activistis into it so that the next gen of the Net isn’t designed purely by American techies.]” – is actually hogwash (imho)

    Have you ever tried running a tigerbox/sniffer for v6 packets over the v4 protocal suite ?? Its will take you many moons. Why ?? B’cos “Internet fulfilled its design spec by delivering the (cut virus) (paste ip packet) to all machines at maximum speed. “. A packet dropped with ttl dead is resent on next available hop !!

    the americans don’t know how to police this paradigm. Therefore obtain buyin/ opt in from “other” countries and the like of cisco and msft !!

    Sorry, I should have blogged my comments and linked into you. But I”ll scratchpad this link !!

Web Joho only

Comments (RSS).  RSS icon