Practical Network-Wide Packet Behavior Identification by AP Classifier

0
426
Practical Network-Wide Packet Behavior Identification by AP Classifier

Practical Network-Wide Packet Behavior Identification by AP Classifier

Abstract

Practical network-wide packet behavior Identifying the network-wide forwarding behaviors of a packet is essential for many network management applications.projects on Practical Network-Wider Including verification of rules, enforcement of policies, detection of attacks, traffic engineering and location of faults.Current tools that can perform identification of packet behavior either incur significant time and memory costs or do not support updates in real time.
 

Experiments using the data plane network state of two real networks show that AP Classifier’s processing speed is at least one order of magnitude faster than existing tools. In addition, AP Classifier uses very small memory and is able to support real-time updates.

Introduction

Practical Network-Wide Packet Behavior Identification by AP Classifier All flow packets have the same forwarding behaviors in a network projects on Practical Network-Wider (also known as the behaviors of the flow) when there is no updating of the data plane.

Network-wide packet behavior identification is a control plane function that discovers the actual forwarding behaviors of the packets in a flow (or a set of flows) including their forwarding paths, where they stop or drop, and which boxes they traverse, by analyzing the network state in the data plane. In the following situations, packet behavior identification is necessary for SDN management.

Conclusion

We propose AP Classifier for the identification of network-wide packet behavior that many important application network management can use. We design algorithms for building the AP Tree for a network to rapidly classify a packet into an atomic predicate.

Projects reports on Practical Network Each atomic predicate represents a set of packets network-wide forwarding behaviors.Experimental results using two real networks datasets show that the proposed AP Tree construction algorithm can optimize the average leaf node depth.AP Classifier can handle millions of packet queries per second. The speed is at least one order of magnitude faster than existing tools.

In addition, it uses only a few MBs of memory in Practical Network-Wide Packet Behavior Identification by AP Classifier. It can be updated in real time and is robust when the data plane changes dynamically.