This event has ended. View the official site or create your own event → Check it out
This event has ended. Create your own
View analytic
Saturday, September 26 • 9:00am - 9:32am
Adding Enhanced Services to the Internet: Lessons for History

Sign up or log in to save this to your schedule and see who's attending!

Paper Link

In this paper, the authors draw on over 20 years of personal involvement in the design and specification of enhanced services on the Internet (Quallity of Service or QoS) to put the current debates into context, and dispel some confusion that swirls around service differentiation. This paper describes the twenty-year failure to get QoS capabilities deployed on the public Internet -- the technical, economic and regulatory barriers, and implications of this failure for the future.

This paper has four parts. The first is a historical perspective, drawing on our own early publications. One paper, from 1994, is a time-capsule of the state of the Internet in 1993. It describes how severe congestion on the NSFnet backbone in the preceding decade compelled its engineers to implement traffic prioritization to support interactive services such as remote login. Another paper, from 1992, framed the architecture of enhanced services within the IP protocol suite -- ideas that were refined, standardized and implemented over the next decade. But nowhere on the global Internet are these services found today.

The next section of the paper reviews the reasons behind this failure, which were mostly unrelated to technology; indeed, technical mechanisms for QoS were deployed in private IP-based networks. Rather, the obstacles to global deployment of QoS mechanisms were coordination failures among ISPs. Although the Internet has a standards body (the IETF) to resolve technical issues, it lacks any similar forum to discuss business issues such as how to allocate revenues among ISPs from enhanced services. ISPs feared such discussions in the mid-90s as risking anti-trust scrutiny. Thus, lacking a way to negotiate the business implications of QoS, it was considered a cost rather than a potential source of revenue.

The third section examines the context of current regulatory resistance to enhanced services -- why the fears of abuse seem to trump potential benefits. We explain why the term "network neutrality" does not capture the desired outcome, which is to prevent abusive behavior. Our paper introduces the basic technical issues that shape QoS implementations, and how they can affect applications running over the Internet.

The fourth section discusses implications of this reality for the future. As IP-based networking subsumes all previous critical communications infrastructure, there is increasing interest in how to satisfy societal requirements for reliability, ubiquity, and resiliency of services in times of crisis. We argue that this last requirement requires prioritizing certain services at times across multi-provider networks. Historical failure of such coordination on the Internet has motivated a shift that will crucially alter the economic and regulatory landscape: a few large ISPs are building private interconnected IP-based networks, using the IP technology but distinct from the public Internet. These private IP networks will allow those ISPs operating them to develop new business models, support enhanced services, and vertically integrate infrastructure and applications.

These networks are a natural industry response to the nation's need for a stable network infrastructure, but they introduce new regulatory concerns. First, this shift leaves the public Internet as a place for games, spam and social play, and perhaps starved for capital investment. Second, the new networks constitute a "shadow" activity, serving a role previously served by a regulated sector that is no longer regulated. Without regulation, these activities may carry substantial systemic risk, amplified in this case by gaps across different bodies of law, which hinder policymakers' ability to respond to problems. In light of these and other prevailing risks of the evolving Internet, we offer several recommendations based on lessons learned from unrealistic assumptions of two decades ago.


Saturday September 26, 2015 9:00am - 9:32am
GMUSL - Room 225

Attendees (26)