Schedule | Location | Parking
# # #
"The Policy Implications of End-to-End," a Stanford Law School workshop, will be held during the day of Friday, December 1, 2000.
The workshop will run from 8:30 am to 5:00 pm:
|8:30 - 9:00 am
|noon - 1:00 pm
||Conclusion of program at Stanford
In the morning, we will explore the values inherent in end-to-end architectures. To this end, we will analyze the implications of an end-to-end design and the effects of deviating from it in the context of particular applications and new developments in network architecture. Four examples will be considered in detail: Voice over IP, network security, quality of service and content caching.
Voice over IP
Two different architectures for providing telephone services over the Internet are currently considered by the Internet Engineering Task Force. One proposal implements the end-to-end principle; the other constitutes a more centralized solution.
Today, network security measures such as firewalls insert points of control within the network. Other proposals such as Internet Security, an architecture providing a whole range of security functions, are more consistent with end-to-end design considerations.
Quality of Service
Increasing network congestion and the requirements of new real-time applications, such as streaming video or telephone over the Internet, have led to proposals that introduce different classes of data transport in the Internet to accommodate the need for a guaranteed quality of data transport. These Quality of Service architectures, however, require a significant amount of intelligent functionality within the network's core, thus deviating from end-to-end design principles.
In order to enable faster access to content on the Internet, content is increasingly cached on servers close to the end-user. This development, however, leads to new requirements in application development, because pure end-to-end applications have to be modified to accommodate the existence of these intermediate servers.
For each of these examples, we will explore the costs and benefits of the different architectural solutions and identify their implications for innovation, competition and end users. Drawing upon this analysis, we will then derive ways to systematically assess the value of end-to-end designs.
The afternoon discussion will focus on two specific types of computer networks: broadband cable networks and wireless networks.
Broadband Cable Networks
Discussion will include the following questions: What is the relationship between end-to-end arguments and open access? Does the existence of facilities-based competition remove the need for open access? What are the implications of open and closed access architectures for innovation, competition and end-users?
We will consider the implications of end-to-end arguments in the context of handset autonomy. We will also discuss new wireless access technologies such as spread spectrum that enable open access to computer networks based on a distributed architecture and will compare them to traditional solutions of allocating spectrum that lead to more centralized access architectures.
The workshop will take place on the Stanford Campus at the Center for Educational Research at Stanford (CERAS). The auditorium, where the workshop will convene, is Room 112.
Coffee, breakfast and boxed lunch will be served in the 1st floor lobby of CERAS.
Parking at Stanford is limited. Day-parking passes may be purchased; alternatively, there are several metered lots on campus. However, we recommend leaving your car at your hotel and taking the Marguerite, Stanford's free shuttle bus system, or one of the free shuttle buses that many hotels operate.