A network processor for a learning based routing protocol

Recently, cognitive packet network (CPN) is proposed as an alternative to the IP based network architectures and shows similarity with the discrete active networks. In CPN, there is no routing table, instead reinforcement learning (random neural networks) is used to route packets. CPN routes packets based on QoS, using measurements that are constantly collected by packets and deposited in mailboxes at routers. The applicability of the CPN concept has been demonstrated through several software implementations. However, higher data traffic and increasing packet processing demands require the implementation of this new network architecture in hardware. In this paper, we present a network processor architecture, which supports this learning based protocol.