On-demand routing is very robust to a fast changing topology but o?®ers
little in guaranteeing a complete path from the source to the destination.
367
Fernand S. Cohen, Joshua Goldberg, and Jaudelice C. de Oliveira
3 Routing in Sensor Networks
There are several ways to classify the di?®erent approaches to routing in sensor
networks. In this section, four separate classes are described: data-centric,
hierarchy-centric, location-centric and trajectory-centric routing.
3.1 Data-Centric Routing
A sensor network??™s ultimate purpose is to sense data and relay it to the sink.
Unlike end-to-end routing problems faced in ad hoc wireless networks, sensor
networks need to aggregate several sources of information into energy e?±cient
transmissions [27]. In data-centric routing protocols, data from several sources
can be combined en route to the sink. This reduces the size of the total data,
the number of transmissions, and thus the total energy consumed.
Data-centric routing di?®ers from the traditional address-based routing
where routes are created between nodes that are addressable. In data-centric
protocols, the sink sends queries to request data, and therefore attributebased
naming is used to specify data characteristics. Data-centric protocols
need to optimize the aggregation of data along the route instead of finding
optimal routes using typical routing metrics. Data-centric protocols are very
useful in situations where data redundancy is present.
Pages:
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587