GIANLUCA PANNOZZO MASTER’S THESIS
Title: Comparison between Generic
Aggregation Protocol (GAP) and Gossip-Based Aggregation
Protocol.
Aggregation
refers to a set of functions that provide global information about a distributed
system. These functions operate on numeric values distributed over the system
and can be used to count network size, determine extreme values, compute averages, products or sums.
In the last years, especially
with the development of wireless sensor networks, data aggregation has received
many attentions
by the researchers. A fundamental challenge in the design of WSNs is to maximize their lifetimes since sensor nodes are
equipped with limited lifetime battery. Data aggregation and in-network
processing techniques are investigated as efficient approaches to
achieve significant energy savings by combining data arriving from different
sensor nodes at some aggregation points en route, eliminating redundancy, and
minimizing the number of transmission before forwarding data to management
station.
GAP is an aggregation protocol
developed by Mads Dam and Rolf Stadler at KTH. This protocol computes Aggregation functions
building a spanning tree over the network, and has shown to provide good
performances in terms of time of convergence, robustness and scalability. The
main drawback of GAP is the large number of messages
exchanged.
This master’s project is
divided into two parts. The main goals of the first part
are:
·
To develop a new
version of GAP in order to reduce the load over the
network.
·
To implement a
Gossip Based Aggregation Protocol.
·
To compare performances of
protocols above. Simulations are carried out on ISP topology, using a simulator
called SIMPSON.
The topic of the second part
is TCA-GAP, where TCA stands for Threshold Crossing Alert. TCA-GAP is a
new protocol developed for
computing TCAs
in a scalable and robust manner. An alert alarm is raised
whenever a management variable exceeds an upper threshold, and is cleared when
the variable decreases below a lower threshold.
e-mail: g.pannozzo@tin.it