AbstractAnycast is a one‐to‐any bidirectional transmission scheme that can be used to provide various service‐oriented applications. However, traditional anycast cannot support some distributed applications that need to locate multiple servers such as network time protocols and threshold cryptography. Although there are some extended anycast schemes that can support the discovery of k services, they may cause high control overhead, low satisfaction ratio, or high searching latency when there are more service requesters and providers. The main reasons for these shortcomings are unknown service information and limited hierarchical structure. This work proposes an efficient anycast scheme, called AnyKast, for discovering k services in cluster‐based mobile ad hoc networks. In the AnyKast scheme, an anycast tree based on the clustering and virtual backbone is established to reduce unnecessary message transmission. In the anycast tree, anycast clusterheads disseminate their information to certain specific nodes that take the service information and perform the service selection task instead of blindly searching. Furthermore, to decrease the cost of service information collection and maintenance, scope flooding is used to limit the information transmission, and information piggybacking and periodical inquiry mechanism are used to increase the service information accuracy. The experimental results demonstrate that our scheme can effectively discover services, reduce request and reply message control overhead, and lower the searching latency. Copyright © 2010 John Wiley & Sons, Ltd.
Read full abstract