Research on the Buffer Management
Algorithm in DTN
Fengju Liu
Inner Mongolia University
Hohhot China
Xiangyu Bai *
Inner Mongolia University
Hohhot China
bxy@imu.edu.cn
Abstract—In Delay and Disruption Tolerant Networks (DTN),
because of its intermittent connectivity features, resulting in
packet transmission process cannot ensure end to end path,
usually employ a "store - carry - forwards" approach to forward
the message. Currently the key method to solve information
exchange in DTN is select forwarding nodes effectively. Made the
information deliver to the destination successfully within a short
time and reduce resource overhead. At the same time, the load of
selected nodes will be increased correspondingly and the
requirement to the buffer also increased, then the management of
the buffer become particularly important. The buffer
management not only includes the message dropping, but also
includes message forwarding. In this paper, we mainly analysis of
the existing buffer management algorithm, and also analysis
some social attributes in favor of selecting forwarding nodes in
DTN . We put forward the buffer management algorithm based
on social attributes in DTN in future work.
Keywords—DTN; forwarding nodes; buffer management;
social attributes
I. INTRODUCTION
DTN [1] was originally developed by KevinFall in ICIR
meeting for solving intermittent connection problems caused
by limited energy, transmission media or distance. DTN is
characterized by long delay, limited node resources,
intermittent connectivity, and its main application scenarios
have continued to increase such as Vehicular Networks,
Underwater Sensor Networks, Pocket Switched Networks,
Vehicular Networks, Military Communications and so on.
Compared with traditional five layers Internet protocol
stack, DTN added a Bundle layer on the transport layer,
forming an overly network architecture. Bundle layer operating
mode is "store - carry - Forward" approach, using
retransmission mechanism with custody ways. A necessary
condition for the use of custody transfer is persistent storage,
while using custody transfer mechanism also brings special
challenges for DTN routing, i.e. the management and use of the
limited storage resources. we can see from Fig.1 shows the
bundle layer entire process of data processing, when there is
insufficient storage space will be very likely to cause a node to
discard the data. So have a good buffer management
algorithms is critical important.
For traditional Delay tolerant networks, the nodes are
considered random walk, with unpredictability, cannot form a
stable end-to-end path, while the Delay tolerant networks with
social attribute, the nodes are considered carried by human,
which move presents a sociological characteristics, not
unpredictable. Based on the results of the analysis of social
networks, researchers have proposed a variety of social
property according to the mobile nodes, i.e. community,
centrality, similarity, selfishness and so on. Of course, these
sociology features can also be applied to DTN.
Fig.1 bundle layer entire process of data processing
Now the key to solve the effective exchange of
information in mobile DTN is to select effectively forwarding
node, so that the information can successfully delivery to the
destination node as possible in a short time and reduce resource
overhead. However, at the same time the load of selected relay
nodes will increase accordingly. If the buffer space is large
enough, the data will be very smoothly forward to destination,
and also can improve the transmission performance. If the
buffer is limited, it is easy to cause the accumulation of
information which lead to relay nodes” congestion, severe can
lead to package loss, if we can manage such nodes reasonable
i.e. forwarding and discarding packages in cache timely, and
ensure the buffer resource can be used Reasonable and
effectively, which will improve the p erformance of the network,
2015 2nd International Conference on Information Science and Control Engineering
978-1-4673-6850-6/15 $31.00 © 2015 IEEE
DOI 10.1109/ICISCE.2015.103
442
2015 2nd International Conference on Information Science and Control Engineering
978-1-4673-6850-6/15 $31.00 © 2015 IEEE
DOI 10.1109/ICISCE.2015.103
442