INK: implicit neighbor knowledge routing in ad hoc networks

Flooding, a common operation providing useful functions in MANET (unicast or multicast routing), could cause redundant rebroadcasts, which in result leads to the broadcast storm problem. Several flooding schemes have been proposed to solve the broadcast storm problem, but none have been regarded as the satisfying solution. Some flooding schemes use the simple probabilistic method while others use neighbor knowledge for efficient flooding. Neighbor knowledge methods use periodic hello packets to acquire neighbor information that generates additional overhead to the wireless network. We propose the implicit neighbor knowledge (INK) routing scheme, which aims to reduce the flooding overhead by using the simple probability based method. INK routing scheme uses a variable probability defined by its neighbor size. INK tries to acquire neighbor knowledge without the usage of any periodic hello packets by observing flooded packets. The proposed scheme keeps up the reachability of blind flooding while maintaining the simplicity of probability based schemes.