The ability to categories the packets into flows in an Internet router is termed as Packet Classification. The packets with same source & destination addresses form a flow which follows a predefined rule and are processed in a similar manner. The packet classification is needed for many sophisticated value added services, such as QoS, load balancing, traffic accounting, etc. Various approaches to packet classification have been studied in the literature with accompanying theoretical bound. In this paper, we present an algorithmic framework for solving the packet classification problem practically. We propose and study a novel approach to packet classification using Multistage classifier Compaction Scheme MCS. Besides high performance, our algorithm preserves average search time and reasonable storage requirements. To evaluate our algorithm, we have developed realistic model of large scale rule bases and use them to drive extensive experimentation. The result shows that the algorithm is very efficient as compared to other techniques in real networks.