BibTeX Entry

  author	= {Erramilli, Vijay and Crovella, Mark},
  title		= {Forwarding in Opportunistic Networks with Resource Constraints},
  booktitle	= {Proceedings of the Fourth ACM Workshop on Challenged Networks (CHANTS 08)},
  location	= {San Francisco, CA},
  month		= sep,
  year		= {2008},
  URL		= {},
  abstract	= {Effective forwarding in mobile opportunistic networks is a challenge, given the unpredictable mobility of nodes, short contact durations between nodes, wireless interference and limited buffer sizes. Most forwarding algorithms aim at decreasing costs (relative to flooding the network) by forwarding only to nodes which are likely to be good relays. While it is non-trivial to decide if an encountered node is a good relay or not at the moment of encounter, it is harder still to prioritize which messages to transmit under the presence of short contact durations and which messages to drop when buffers become full. The main objective of this paper is to study different message prioritization schemes using real measurements. Such schemes can be broadly divided into two categories - schemes which do not use any network information, and schemes which do. Examples of the former set of schemes include FIFO/LIFO etc. For the latter set of schemes, there is a key design choice: On one hand, we have the following scheme: when a forwarding opportunity presents itself, assign high priorities to messages which are relatively close to their intended destination. On the other hand, we can assign high priorities to messages which are farther away from their destination than closer messages. In order to decide if messages are close to their destination or not, we have to rely on a forwarding algorithm. For this, we use delegation forwarding schemes which have been shown to be efficient in terms of cost incurred in the network. We develop a new set of prioritization schemes based on delegation schemes. We consider these schemes in our empirical study.}