As discussed in the previous lectures, ossification of the Internet transport-layer architecture is a significant barrier to innovation of the Internet. Such innovation is desirable for many reasons. Current applications often need to implement their own mechanisms to receive the transport service they need, but many do not have the breadth of adapting to all possible network characteristics. An updated transport architecture can do much to make the Internet more flexible and extensible. New ground-breaking services often require different or updated transport protocols, could benefit from better signaling between application and network, or desire a more flexible choice of which network path is used for which traffic.

This lecture introduces a new transport architecture,  a New, Evolutive API and Transport-Layer Architecture (NEAT), which targets the ossification of the Internet transport-layer architecture, and which is designed to offer a flexible and evolvable transport system. The design of NEAT is based on the transport services concept and NEAT is one of the systems that has served as input for the ongoing standardization of the Transport Services (TAPS) Architecture which was introduced in the previous lecture. Networked applications interface NEAT through an enhanced API that effectively decouples them from the operation of the transport protocols and the network features being used. In particular, applications provide NEAT with information about their traffic requirements. Then, on the basis of these requirements, pre-specified policies, and measured network conditions, NEAT establishes and configures appropriate connections.

Course lecture: NEAT

Related materials

The first paper provides a fairly comprehensive overview of the NEAT transport architecture, including the rationale behind NEAT; its conceptual architecture; its protocol-agnostic and evolvable transport API; and, not least, how the concepts behind NEAT are currently being standardized in the IETF TAPS working group. The remaining two papers demonstrate how NEAT could benefit applications in multi-access environments, and how it could reduce the download times for a web browser.

  • N. Khademi et al., “NEAT: A Platform- and Protocol-Independent Internet Transport API,” in IEEE Communications Magazine, vol. 55, no. 6, pp. 46-54, June 2017. [pdf available here] (Instructions for how to access)
  • P. Hurtig et al., 2017. A NEAT Approach to Mobile Communication. In Proceedings of the Workshop on Mobility in the Evolving Internet Architecture (MobiArch ’17). ACM, New York, NY, USA, 7-12. [pdf available here] (Instructions for how to access)
  • F. Weinrank et al., 2017. A NEAT Way to Browse the Web. In Proceedings of the Applied Networking Research Workshop (ANRW ’17). ACM, Prague, Czech Republic. [pdf available here]

Additional readings

The two additional readings are “technical” papers that provide more technical details on specific  aspects of NEAT. The first paper focuses on one of the more salient features of NEAT, Happy Eyeballs — it explains how the Happy Eyeballs mechanism works and evaluates the cost and benefits of using this mechanism. The second paper details the design of the NEAT policy manager and how it can interact with a SDN controller.

  • G. Papastergiou et al., 2016. On the Cost of Using Happy Eyeballs for Transport Protocol Selection. In Proceedings of the Applied Networking Research Workshop (ANRW ’16). ACM, Berlin, Germany. [pdf available here]
  • Z. Bozakov et al., “A NEAT framework for enhanced end-host integration in SDN environments,” 2017 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN), Berlin, 2017, pp. 1-7. [pdf available here] (Instructions for how to access)