Comments on: Joe Lieberman on End to End Blog, news, books Thu, 12 Oct 2017 05:34:00 +0000 hourly 1 By: tonyl Sat, 11 Oct 2003 22:38:57 +0000 Electronic voting machines video:

By: Richard Bennett Tue, 15 Jul 2003 09:09:03 +0000 I don’t doubt that Lessig means well, but he frankly doesn’t know what he’s talking about. There is a legitimate, if obscure, fear in some quarters that ISPs may someday censor specific types of content, either in their customer’s interest or in their own economic interests. Porn filters, for example, discriminate based on content, but many customers would consider this a value-added service and it’s not worth getting excited about.

But let’s take it a step further, and suppose that an ISP filters video packets, ostensibly because it wants to control your video experience through its ownership of your cable TV franchise. This would be a bad thing, of course, and I don’t argue otherwise.

But the question we have to ask as network architects is whether there’s any relationship between the Internet’s present or future architecture and this sort of censorship, and the answer to that is clearly no. Video packets are easy to identify on the net because they’re carried by a limited set of protocols and clearly marked; a censor doesn’t care whether the marking is at the IP layer or at the RTP layer or at the UDP layer; they each have a signature, and unless they’re encrypted, they can be found.

Now the question has to be asked as to whether the Internet’s current architecture can hope to compete with cable TV and DBS as a practical alternative for carrying audio and video data, and whether this should be a goal. In the early days of IP, it clearly wasn’t a goal and therefore an architecture was developed that blocked transport layer access to the isochronous services in the data link and medium access control protocols that would make it practical. This architecture now has the effect of keeping the voice and data networks separate, to the advantage of telcos and cablecos who would like to bill you extra for providing voice and video services.

So far from advocating an architecture that frees the consumer from the big media and telephone companies, the end-to-end cargo cultists are promoting the exact thing that keeps them dependent, and they do so out of ignorance of the technical issues in network architecture.

To put it simply, you wouldn’t trust me to explain constitutional law to you, so why would you trust someone with Lessig’s background to explain my business, network architecture, to you?

By: Matt Lupfer Tue, 15 Jul 2003 03:05:10 +0000 Richard –

In his book _The Future of Ideas_, Prof. Lessig applauds the positive externalizes of end-to-end protocols and warns that if the Internet were to discriminate against packets, the quality and quantity of net-enabled innovation would suffer. To summarize (although I doubt it is necessary here), he writes that by allowing network operators to make QoS decisions, we would be opening a kind of Pandora’s Box that would allow more control to seep into the infrastructure of the Internet, thus nullifying the gains made from a network that is ignorant of what’s going on at the Application Layer. A “dumb” network, Prof. Lessig writes, is fertile ground for the development of new protocols and applications.

In your comment (if I understand it correctly), you suggested that IP is flawed because its inability to discriminate and make routing decisions based on the requirements of the application. How would you respond to Lessig’s criticisms of “smart” (or perhaps more accurately “optimized”) networks? Can the technological requirements of optimizing the Internet for multicast/broadcast be met without centralizing control?

I don’t mean to sound like a Lessig disciple, as I have some serious questions for him that will have to wait for another day, I’m afraid. However, I would be very interested to read your response.

By: Richard Bennett Thu, 29 May 2003 17:33:48 +0000 The link you provide for the Lieberman paper – – is a Not Found.

Lieberman’s understanding of the Internet is on par with that of the authors of the “World of Ends” paper, which is to say, not much. The IP protocol is flawed by not providing for real-time, sequenced delivery of packets, or for real multicast, which hampers video and audio and makes “many-to-many” less than it appears.

I’ve seen that the people touting end-to-end as an ideal architecture (rather than a momentary compromise appropriate in 1982 but not today) don’t understand these issues.

Oddly, hamstringing the Internet behind this out-dated End-to-End design means that it can never effectively counter broadcast, which is something opponents of streamlined media should want.

By: anon Wed, 28 May 2003 23:26:55 +0000 I can’t be sure but I believe I read an article in Wired (or Fast Company or Business 2.0) that was heavily plagiarized by Lieberman when he wrote this… Off to find it now…

By: Donna Wentworth Wed, 28 May 2003 16:17:45 +0000 Oh, thank heaven. This is terrific.