Effective buffer management and scheduling of bundles in delay tolerant networks with finite buffers

Delay tolerant networks (DTNs) are a class of emerging networks that experience frequent and long-duration partitions. These networks have a variety of applications in situations such as crisis environments and deep-space communication. In this paper, the problem of routing in DTNs is studied. While...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Fathima, G., Wahidabanu, R.S.D.
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Delay tolerant networks (DTNs) are a class of emerging networks that experience frequent and long-duration partitions. These networks have a variety of applications in situations such as crisis environments and deep-space communication. In this paper, the problem of routing in DTNs is studied. While routing in the Internet and mobile ad hoc networks has been studied extensively, due to the unique characteristic of frequent partitioning in DTNs, traditional ad hoc routing protocols such as AODV and DSR fail to establish route. Several algorithms are coming up to support routing in DTNs. In this paper, the issues of using epidemic routing in DTNs have been focused. The goals of epidemic routing are to: i) maximize message delivery rate, ii) minimize delivery latency. According to the previous work, to ensure eventual delivery of all messages, the buffer size on at least a subset of nodes must be roughly equal to the expected number of messages in transit at any given time. i.e. It has the constraint of infinite buffers at each node. But this is not always valid in realistic scenario. The tradeoff between buffer size and message delivery are observed and understood that maintaining finite buffers will not affect the delivery ratio significantly. But when finite buffers are maintained, the problem of what should be done when buffer is full arises. So some messages have to be flushed from the buffers before delivery to give room for new messages. In this paper, an appropriate buffer management scheme based on the expiration time has been proposed. Our approach is to maintain separate queue for each class of service and use weighted fair queuing algorithms to schedule bundles based on the class of service.