ChoiceNet

An Economy Plane for the Internet

Overview

The Internet has been a key enabling technology for many new distributed applications and services. However, the deployment of new protocols and services in the Internet infrastructure itself has been sluggish, especially where economic incentives for network providers are unclear. In this project, we seek to develop an "economy plane" for the Internet that enables network providers to offer new network-based services (QoS, storage, etc.) for sale to customers. The explicit connection between economic relationships and network services across various time scales enables users to select among service alternatives. The resulting competition among network service providers will lead to overall better technological solutions and more competitive prices.

This project is one of five NSF-sponsored Future Internet Architecture (FIA) projects.

Vision

The goal of ChoiceNet is to enable choices and the associated economic relationships between entities in the network. ChoiceNet makes it possible for network service providers to compete for customers and be rewarded for quality and innovation. In today’s network, money enters the network ecosystem only around the edges: consumers (individuals or enterprises) pay access providers for Internet service. Most access providers, in turn, pay other ISPs to carry their traffic to/from the rest of the Internet. Indeed, most end-to-end traffic in the Internet traverses at least three distinct service
providers.

Thus, in today’s Internet (i) no single provider controls all end-to-end paths; (ii) money flow between providers is outside the architecture and by necessity changes slowly; and (iii) traffic flow is constrained at the granularity of providers to follow the money flow. The result is that transit providers have neither means nor incentive to compete via new service offerings, and consumers have essentially no control over the service they receive or its quality. A central thesis of ChoiceNet is that enabling money flow to follow traffic flow (instead of vice versa), coupled with greater support for choice among end-users, should lead to increased provider competition and more innovation.

ChoiceNet’s economy plane aims to give assurances to providers that they can compete for customers and be compensated for the services they render. At the same time, ChoiceNet provides users with the ability to select from a set of offerings and combine them to form complex services, thereby separating services that are currently entangled in the current Internet. Key to such an architecture is the ability to market services and then form or dissolve business relationships on (potentially small) time scales. Moreover, ChoiceNet must enable providers (and consumers) to prove (or verify) that the contracted service was rendered as promised.

Example

To illustrate the vision for ChoiceNet with a specific example, consider how choice can motivate the deployment of innovative new services and protocols for streaming video content. In the current Internet, there are a number of different content providers that offer video streaming services (e.g., Netflix, Hulu, Amazon, etc.). While users can choose a video content provider, users cannot choose how the network handles the streaming traffic produced by the provider as it traverses the network. As a result, users can only hope that they receive the content with enough quality of service to have an enjoyable experience. Content providers also have little or no control over the network when sending their content and instead focus on coding and quality adaptation techniques to adjust to changing network conditions.

In a ChoiceNet-enabled Internet, we envision that users will explicitly select network services from a marketplace of competing service offerings. Every aspect of the network can be marketed and sold as a service. Moreover, it is possible to combine services, such as specific network paths, in-network storage, packet forwarding prioritization, etc., into a tailored movie-watching package. These packages can be crafted by knowledgeable users, their applications, or, more likely, by service providers. Users can choose among different packages, experience their quality, and decide which package they want to continue using and paying for.

While this scenario requires significant changes in how economic relationships are established between customers and providers, there is also need for technological change. Services need to be general, so that they can be combined into useful packages. Also, the network infrastructure itself needs to be diverse enough to offer a variety of choices at the network layer. While the Internet offers a wide range of end-system services, the set of end-to-end network/transport layer services available is rather small. However, recent advances in software defined networking, programmable networks, and cloud computing make it possible to offer such alternative services.

Not only should ChoiceNet support alternatives for consumers to choose from and purchase, but it also needs to support a variety of economic relationships. Providers may bundle and resell services offered by others, adding value in the process; in doing so they act as both customers and providers. For example, today’s mobile virtual network operators (MVNOs) provide cellular network access to users through short-term service contracts, but do not operate their own infrastructure. Instead they resell network access from other providers. Similarly, crowdroaming enables net work access to participants by sharing access points. These current-day examples, however, lack a general architecture (or in some cases specific economic models), which is what ChoiceNet aims to provide.

News

Papers

The ChoiceNet architecture is described in these papers:

Additional papers that address specific issues within ChoiceNet are here:

Team

The ChoiceNet project is lead by four institutions: the University of Massachusetts Amherst, the University of Kentucky, North Carolina State University, and the Univeristy of North Carolina.

The principle investigators and co-principle investigators on the project are:

Graduate students on the project are:

Code

There is no publicly available implementation of ChoiceNet, yet.

Developer link

Sponsor

The project is supported by the National Science Foundation under Grant Nos. 1111040, 1111088, 1111256, 1111276.