Elon University
The prediction, in brief:

How might we go about implementing an expected capacity scheme? … This mechanism has two parts: traffic flagging, which occurs in a traffic meter at access points, and congestion management, which occurs at switches and routers where packet queues may form due to congestion.

Predictor: Clark, David D.

Prediction, in context:

David D. Clark stated the following in a presentation at the MIT workshop on Internet Economics in March 1995: ”How might we go about implementing an expected capacity scheme? As an example, here is a specific mechanism. In fact, this proposal is overly simple, and does not address certain issues that may be very important. These are identified later. But this discussion will provide a starting point. This mechanism has two parts: traffic flagging, which occurs in a traffic meter at access points, and congestion management, which occurs at switches and routers where packet queues may form due to congestion. At the access point for the user, there will be some contract as to the expected capacity that will be provided. Based on this contract, the traffic meter examines the incoming stream of packets and tags each packet as being either inside (in) or outside (out) of the envelope of the expected capacity. There is no traffic shaping – no queuing or dropping. The only effect is a tag in each packet. To implement a pure expected capacity scheme, this tagging always occurs, whether or not there is congestion anywhere in the network.”

Biography:

David D. Clark was a senior research scientist at MIT’s Laboratory for Computer Science. (Pioneer/Originator.)

Date of prediction: March 1, 1995

Topic of prediction: Information Infrastructure

Subtopic: Pipeline/Switching/Hardware

Name of publication: The Journal of Electronic Publishing

Title, headline, chapter name: A Model for Cost Allocation and Pricing in the Internet

Quote Type: Direct quote

Page number or URL of document at time of study:
http://www.press.umich.edu/jep/works/ClarkModel.html

This data was logged into the Elon/Pew Predictions Database by: Catalfumo, Cara J.