We compare them in energy
e?±ciency, control message overhead, cluster distribution, i.e., the average cluster
size and the variation of cluster sizes, as well as stability in the presence
of node mobility.
Energy-e?±ciency is an important metric. Sensor nodes are typically densely
deployed and highly energy-constrained. Further, recharging or replacing batteries
on hundreds or thousands of nodes in possibly hostile environment is
infeasible ([8], [9] and [1]). The key challenge in such a system is conserving
sensor energy, thus maximizing network lifetime. Among many other activities,
to gather the sensory data periodically and transmit it to the base station
is the most energy consuming. Clustering plays a significant role in energy
conservation.
Message overhead is another important metric for comparing clustering
schemes, as it measures the scalability of a clustering scheme, the degree to
which it will function in a congested or low-bandwidth environment, and its
e?±ciency in terms of consuming battery power. We evaluated the number of
control messages used by each scheme during cluster formations.
In sensor networks, the size of clusters and the number of clusterheads as
well as the stability of clusters are of considerable importance to the higher
layer protocols [11]. Since each clusterhead needs to keep track of the routing
information of other clusterheads in the system, the larger the number of clusterheads,
the larger will be the overhead in sharing the routing information.
Pages:
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558