Vehicular Sensor Networks: General Aspects and Implementation Issues - Application and Multidisciplinary Aspects of Wireless Sensor Networks - page 211

Fig. 10.2 The push-based dissemination approach
Fig. 10.3 The pull-based dissemination approach
in the same data making an inefficient usage of network resources. The push-
based approach is optimized for many receivers and few senders, Fig. 10.2 .
The idea of pull-based dissemination is based on the request-response model
and it is useful for user-specific data, e.g., webpage request. Motivated by applica-
tion performance requirements, algorithms for dissemination are developed to
match them to application requests like one-phase pull (an algorithm optimized for
many senders and few receivers) and two-phase pull . Due to lots of cross traffic,
this approach is characterized with contentions, collisions, and interference. The
pull-based dissemination approach is given in Fig. 10.3 .
The flooding approach is based on broadcasting the generated and received data
to neighbors and usually everyone participates in the dissemination. This method is
good for delay sensitive applications. However, the problem of broadcast storm
arises. Techniques for avoiding it are based on simple forwarding, which can be
timer based and hop limited, and Map-based/Geographic forwarding, which can be
based on directed flooding or aggregation.
The relaying approach is based on the idea that the relay node forwards the data
to the next hop and so on. The decision on which node to be selected as a next hop
(relay neighbor) is either based on simple forwarding, by selecting the node which
is farthest from the source, or is based on Map-based/Geographic forwarding which
selects the node that is closest to the destination.
The architecture and the communication in VSNs are not necessarily only V2V
or only V2I/I2V. When there is a hybrid architecture, i.e., V2V and V2I/I2V at the
same time, the communication and the delivery of data between vehicles and
among vehicles and RSUs must be improved. Additional data dissemination
approaches can be obtained by combining the basic methods previously mentioned
leading to opportunistic , geographical , peer-to-peer, and cluster-based dissemina-
tion. Their characteristics are briefly elaborated in the following paragraphs.