In this chapter, we consider two specific classes of networks for DSNs: (a) wide-area networks with the requirements of sustained bandwidth and stable control channels, and (b) small-area ad hoc wireless networks of mobile nodes deployed in unstructured areas. To address the first class, we employ the regression-based path computation  and transport stabilization based on stochastic approximation method . Ÿese daemons collect link delay measurements to compute the best paths and use each other to route around high tra«c areas as well as to accumulate bandwidth using multiple paths. For implementing stable control channels, we adopt a source controller that stabilizes a °ow using User Datagram Protocol (UDP). We analytically justify both techniques under fairly general conditions. For the second class of networks, the daemons dynamically track the connectivity changes and exploit the node movements to improve the message delivery ; this is a departure from the Internet-based approaches that treat connectivity changes as undesirable.
|Original language||English (US)|
|Title of host publication||Distributed Sensor Networks|
|Subtitle of host publication||Second Edition: Sensor Networking and Applications|
|Number of pages||23|
|State||Published - Jan 1 2016|
All Science Journal Classification (ASJC) codes
- Computer Science(all)