Smart Packet Forwarding in Router using Classification Of System Traffic with Neural System

Abstract

Router assumes a critical part in arrange directing or organize flows sending, the router carry the packet to one place to other via preeminent way. In traditional method router forward packet to only known predefined path, unaware of importance of packet and packet count total. It diminishes the productivity of router when the way is inaccessible. If router has knowledge on ingress traffic, easy to manage bandwidth allocation when there is heavy traffic. In proposed method Prior knowledge on traffic is determined by neural network traffic classification method. classification methods like port based and signature methods are give less accuracy, comparatively proposed neural network gives more accuracy then the traditional classification methods. Classification result is optimized by fish seminary search algorithm; this classification result is forwarded to router for efficient packet delivery.