An Efficient Identification of Malicious Packet Dropping in Wireless Sensor Networks

In the wireless sensor networks, sensor nodes monitor the environment detect the event of interest, produce data and it involves in forwarding the data towards a sink, which could be a gateway, base station, storage node or querying user. In such hostile environment it lacks physical protection and it is subjected to various attacks. Dropping packets and modifying packets are most common attacks that can be launched by an attacker to disturb the communication in wireless sensor networks. In this paper, we proposed a node categorization algorithm to indentify the nodes which are suspicious droppers/modifiers. Node categorization algorithm performs better, when compared to PNM scheme (Probabilistic Nested Marking Scheme). In our proposed scheme, a routing tree ischolar_mained at the sink is first established. Each packet is encrypted and padded so as to hide the source of the packet. The packet mark is small number of extra bits is added in each packet, such that the sink can recover the source of the packet and then find out dropping ratio associated with every sensor node. The routing tree dynamically changes in each round so that the behaviour of sensor node can be observed in different scenarios. So we are implementing this scheme which is effective in identifying both packet droppers and modifiers. It is being compatible with existing false packet filtering schemes that is, it can be deployed together with the false packet filtering schemes. In addition to identifying intruder filters, it also detects the filter modified packets immediately after the modification.