A packet processor for a learning-based routing protocol

ABSTRACT

As the Internet expands significantly in number of users, servers, routers and other networking products, the IP based network architecture must evolve and change. There are already proposed alternative packet-switched network models that would eliminate some of the problems of IP based networks. Recently proposed Cognitive Packet Networks (CPN) is one of them and it shows similarity with the discrete active networks. CPN uses a neural network model with a reinforcement learning algorithm to find routes. In this paper, we present a packet processor architecture which supports this fast adaptive routing algorithm.

REFERENCES

• 1: Erol Gelenbe, Learning in the recurrent random neural network, Neural Computation, v.5 n.1, p.154-164, Jan. 1993 [doi>10.1162/neco.1993.5.1.154]
• 2: D. Tennenhouse, J. Smith, W. Sincoskie, D. Wetherall, G. Minden, "A survey of active network research", IEEE Communications Magazine, vol. 35, no. 1, pp. 80--86, 1997.
• 3: D. Wetherall, J. Guttag and D. Tennenhouse, "ANTS: A toolkit for building and dynamically deploying network protocols", Proc. IEEE OPENARCH, Apr. 1998.
• 4: Erol Gelenbe , Ricardo Lent , Zhiguang Xu, Design and performance of cognitive packet networks, Performance Evaluation, v.46 n.2-3, p.155-176, October 2001 [doi>10.1016/S0166-5316(01)00042-6]
• 5: Zhiguang Xu , Erol Gelenbe, Design and evaluation of adaptive routing in cognitive packet networks, 2001

INDEX TERMS

Primary Classification:
  B. Hardware
  B.4 INPUT/OUTPUT AND DATA COMMUNICATIONS
      B.4.1 Data Communications Devices

General Terms:
Design

Keywords:
intelligent routing algorithms, network processors, neural networks