Communication Cost Reduction by Data Aggregation: A Survey

14 International Journal for Modern Trends in Science and Technology
Volume: 2 | Issue: 06 | June 2016 | ISSN: 2455-3778IJMTST
Communication Cost Reduction by Data
Aggregation: A Survey
Nithin Kumar1
| Nagarathna2
1M.Tech Scholar, Department of CS&E, P.E.S College of Engineering, Mandya, Karnataka, India
2Associate Professor, Department of CS&E, P.E.S College of Engineering, Mandya, Karnataka, India
Wireless Sensor Networks have gained wide popularity in the recent years for its high-ranking applications
such as remote environment monitoring, target tracking, safety-critical monitoring etc. However Wireless
Sensor Networks face many constraints like low computational power, small storage, and limited energy
resources. One of the important issues in wireless sensor network is to increase the network lifetime to keep
the network operational as long as possible. In this survey paper, we provide a comprehensive review of the
existing literature on techniques and protocols for data aggregation to reduce communication cost and
increase network lifetime in wireless sensor networks.
KEYWORDS: Data Aggregation, Wireless Sensor Network, Communication cost, Energy Efficiency.
Copyright © 2015 International Journal for Modern Trends in Science and Technology
All rights reserved.
I. INTRODUCTION
A wireless sensor network (WSN) is a group of
sensors nodes (transducers) that are deployed
across a geographical area for regulating and
monitoring that area for physical phenomena’s like
sound, temperature, particular events and so on.
Typically a sensor is a small device that consists of
three main components such as, sensing
subsystem which is responsible for data
acquisition from the environment in which sensor
is deployed to work, processing subsystem which is
responsible for local data processing and storage
within that sensor node,wireless communication
system is responsible for data transmission.
Fig 1:- Sensor Network Architecture
On the other hand, a wireless sensor network
should possess long lifetime so that it can fulfill the
requirement of deploying it, so it may take several
days or months or even it may take years together
to fulfill the requirement of deploying it. Therefore
energy conservation is a key constraint in the
design issue of system’s based on wireless sensor
networks.
Experimental measures have shown that
generally communication cost is very high in terms
of energy consumption compared to data
processing [1], is that the communication cost for
transmitting a bit of information is exactly the
same energy required for processing thousands of
instructions in a sensor node [2]. Therefore the
lifetime of a wireless sensor network (WSN) can be
extended by applying some sort of techniques on
two subsystems that reduces the energy
consumption that are, network subsystem in which
energy consumption was taken into account on the
operation’s performed on a sensor node by using
some networking protocols, sensing subsystem in
which energy consumption was taken into account
by reducing the amount and frequency of high
energy expensive samples. The lifetime of WSN can
be extended by applying different techniques [3],
for example energy management protocols are
applied on nodes during data transmission to
ABSTRACT

Communication Cost Reduction by Data Aggregation: A Survey
reduce energy consumption, similarly some nodes
consume more energy even in idle state so for that
sort of energy consumptions, power management
schemes are used for switching off node
components that are not needed for particular
processing.
In this survey paper we majorly focus on
reducing communication cost by data aggregation
and data management, these techniques allows us
to trade-off communication for computational
complexity in any application area and in fact when
compared to communication cost, the local
computation consumes less power.
Data aggregation and data management are the
most important part of network research in which
available resource efficiency, timely delivery of
computational results, accuracy of obtained
results are conflicting goals, and all these mainly
depends on the application so its application
specific. Basically, in data aggregation techniques
consists of different methods to route all packets in
order to combine the data coming from different
sources but combined data will be forwarded into
same destination.
The main aim of this survey paper is to provide
an overview of data aggregation by defining the
main concepts and most important and recent
work in field of reducing communication cost in
wireless sensor networks (WSNs), and on other
hand to recognize and propose directions for future
research in this area. The survey paper is organized
as follows. In Section II we explain general
approaches that are available for reducing energy
consumption in WSNs. In Section III we describe
data aggregation paradigm to classify existing
algorithms. In Section IV we introduce some
network protocols with data aggregation and
classify existing solutions. In Section V finally we
summarize the different data aggregation
approaches discussed throughout the survey
paper and provide the directions and motivations
of future research work in this area.
II. GENERAL APPROACHES TO ENERGY
CONSERVATION
Before representing a high-level classification
of energy conservation proposals, we should
know about network and node-level architecture.
Fig 2: Represents the Architecture of typical Wireless
Sensor Node.
Basically it is composed of four main
components such as, sensing subsystem includes
one or more sensors for data acquisition,
processing subsystem includes a microcontroller
and small amount of memory for local data
processing, radio subsystem for wireless data
transmission, power supply unit which will be
based on specific application and it may include
location search system and mobilizer to change
their location and so on.
The radio subsystem consumes more energy
than the processing subsystem. It has been shown
that transmitting a bit consumes more power than
executing thousands of instruction [2]. The radio
energy consumption is of same order in receiving,
transmitting, and even in idle states, so whenever
energy drops, therefore the radio should be put into
sleep or turned off whenever possible. The sensing
subsystem may consume more energy depending
on applications, it’s an application specific
constraint so it should be taken care.
Fig 3: Hierarchy of Energy Conservation Schemes
Energy
Conservation
Schemes
Duty Cycling Data Driven Mobility
Based
Data
Compression
Data
Aggregation
Data
Prediction

A. Duty cycling is an approach of energy
conservation in which, a sensor is switched to
low power sleep mode, whenever
communication is not required that is, if there
is no data to send/receive. Sensor should be
resumed as soon as new data becomes
available. In this survey paper we won’t
concentrate on these sort of techniques, so
interested readers can refer the papers [4], [5],
[6], [7], [8].
B. Mobility based schemes in which mobile
nodes can be divided into two categories such
as, part of network infrastructure and part of
environment. Whenever the mobility nodes are
part of infrastructure their mobility can be fully
controlled and customized. When nodes are
part of environment they are not controllable.
Finally mobile nodes may follow a mobility
pattern that is neither predictable nor
completely random. In this survey paper we
strictly concentrating on data aggregation to
reduce communication cost, so mobility is not
discussed here, interested readers can refer the
papers [9], [3], [10], [11].
C. Data driven approaches can be classified into
two subclasses such as, data reduction, which
aims to reduce the unneeded samples, and
energy efficient data acquisition is an approach
that mainly concentrates on energy spent by
sensing subsystem. Data compression is a part
of data reduction in data driven approaches in
which the compression can be applied to
reduce the amount of data sent from source
node. Data prediction is a approach of
developing an abstraction of sensed data, i.e. a
model that defines the data that may evolve,
and the model can predict the data sensed by
sensor nodes.
In this survey paper we mainly concentrate on
data aggregation approach of data driven so that
interested readers about data compression and
prediction can refer the following papers for more
details about that [12], [13], [14], [15]. Basically
data aggregation is an technique of aggregating
data (computing average) at intermediate node
between the source node and sink, so that the
amount of data is reduced while transmission, in
this way communication cost can be reduced by
using data aggregation.
III. DATA AGGREGATION PARADIGM
Basic scenario of WSNs is the data collected
from different sensor nodes, which will be made
available in sink node, at sink node, collected data
is analyzed and processed by particular
application. The data produced by different sensors
can be totally processed while being transmitting
towards the sink, e.g., by combining together all
sensor readings related to the same event or any
physical quantity, or by locally processing raw data
before the transmission. Data aggregation
techniques consists of, how data is collected at the
sensor nodes, and how data packets are routed in
the network, and have more impact on energy
conservation and efficiency (e.g., by reducing
transmissions or amount of data to be
transmitted).Data aggregation can be complex
task, since the aggregation algorithms should be
distributed in the network and requires
co-ordination among sensor nodes for better
performance.
Data aggregation is an technique of aggregating
data (computing average) at intermediate node
between the source node and sink, so that the
amount of data is reduced while transmission, in
this way communication cost can be reduced, which
in turn increase’s network lifetime.
There are two common approaches of data
aggregation:
 Data aggregation with size reduction is a
process of fusing and compressing data
obtained from different sources to reduce the
amount of information to be transmitted in the
network. As an example, assume that a node
receives two packets from two different source
nodes containing some readings. Instead of
forwarding the two packets, the sensor may
compute the average of the two readings and
transmitted as a single packet.
 Data aggregation without size reduction is a
process of combining data coming from
different sources into the same data packet
without processing: assume that two packets
carrying different physical quantities, for e.g.,
sound and vibration, these two values cannot
be processed together but they can be sent as a
single data packet, thereby reducing
communication cost and overhead.
The first approach is good one for reducing the
amount of data that may be transmitted over the
network, but it will reduce the accuracy, because

after applying the aggregation operation, it is not
possible to reconstruct the original data
successfully. The second approach, in which the
accuracy of original information is maintained, but
these sort of approaches usually depends on many
factors including the application type, the data
rate, the network properties, and so on.
In data aggregation techniques, there must be
some sort of synchronization between nodes. In
some cases, the bet strategy is that at a given node
is not always to transmit the data as soon as it is
available. Node must wait for information from
neighbor node, this may lead to better data
aggregation and increase in performance. Timing
strategy must be required, especially in some
monitoring applications where sensor node should
report their data within some particular timeslot.
Based on timing strategies involved in data
aggregation, it is been classified as follows:
 Periodic simple aggregation in which, a node
must wait until predefined timeslot to
aggregate all data values that it received, and
then sends a summarized value of all received
data within that time as a single packet.
 Periodic per-hop aggregation it is similar to
periodic simple aggregation, but the only
difference is that the aggregated value is
transmitted as soon as node hears from its
child nodes. This approach requires each root
node to know about its child nodes.
 Periodic per-hop adjusted aggregation in which
the time lot to send aggregated data can be
adjusted based on the position of node in the
tree.
The most important functionality that data
aggregation techniques must provide is the ability
to aggregate the data obtained from different
nodes. Based on aggregation function, the data
aggregation approaches can be classified as shown.
 Lossy and lossless in which aggregation
functions is applied on amount of data
information that is to be transmitted in the
network. The first approach is good one for
reducing the amount of data that may be
transmitted over the network, but it will reduce
the accuracy, because after applying the
aggregation function, it is not possible to
reconstruct the original data successfully. The
second approach, in which the accuracy of
original information is maintained, that is all of
its original readings can be perfectly
reconstructed from their aggregated value at
receiver end.
 Duplicate sensitive and duplicate
insensitive in data aggregation process an
intermediate node may receive multiple
data packets of the same information. In
such case, the same data is considered
multiple times during data aggregation. If
the aggregation function is duplicate
sensitive, the final aggregated result
depends on the number of times the same
value has been taken during aggregation.
Otherwise, the aggregation function is
duplicate insensitive.
IV. NETWORKING PROTOCOLS FOR DATA
AGGREGATION
The concepts that are discussed so far deals
with transmission of data packets to facilitate data
aggregation process of information. Basically, the
idea is to enhance existing routing algorithms to
improve the performance of data aggregation to
reduce communication cost. In this survey paper
we mainly focus on four classes of routing
protocols such as tree-based, cluster-based and
multi-path. In these approaches, the tree based
approach is a classical approach, which consists of
routing algorithm based on tree rooted at the sink
node. In some cases, the tree-based structures are
grouped into particular cluster based on the type of
data gathered in it, to improve efficiency of data
aggregation. In this survey we focus on multi-path
routing, which overcomes the drawbacks of
tree-based routing. Finally, some recent
researcheshave been done to improve the mixture
of tree-based and multi-path routing are called as
hybrid approach to improve efficiency and adaptive
nature of existing routing algorithm.
Fig 4: Different Data Aggregation Approaches
Data
Aggregation
Approaches
Tree-Based
Approach
Cluster-Based
Approach
Multi-Path
Approach

A. Tree-Based Approaches
Classic routing techniques [16], [17] are
usually based on hierarchical organization of
nodes in network. In tree-based approaches, the
data is flowing from source node to sink node,
between them selected intermediate nodes perform
aggregation function on obtained data and then
transmits the aggregated data in preferred
direction. In tree-based approaches, a node may be
selected based on some criteria’s such as its
position within the tree [18], its resources [19], and
processing cost of aggregation [20]. In tree-based
approaches, a spanning tree is constructed rooted
towards the sink node, these constructed
structures is exploited upon generation of queries
by the sink node. This can be done by performing
data aggregation along the aggregation tree by
proceeding level by level from its leaf node to its
root node. These kind of tree-based approaches
have some failures, because WSNs are not always
free from failures. When a data packet is lost due to
some failures in tree at that level E.g., node failure,
may lead to loss of data coming from related sub
tree, so tree-based approaches are suitable in
designing optimized aggregation function to
perform energy management. Finally, a new
approach that is based on tree-based routing is the
construction of connecting dominating sets [21]
which consists of selected number of nodes, they
form a network in which any node can collect data
from any point in network. These connecting
dominating sets construction is recommended for
energy balancing.
1. TAG [22]
The Tiny Aggregation (TAG) is a data centric
approach, it is particularly designed for monitoring
systems which produces similar information
periodically. TAG approach can be classified as
periodic per-hop adjusted aggregation approach.
TAG algorithm can be implemented in two phases
distribution phase where queries are distributed to
sensors, and collection phase where aggregated
sensor values are obtained. TAG is a tree-based
approach which is rooted at sink node, the sink
node broadcasts a message to nodes for the
construction of tree structure, upon receiving that
message every node then re-broadcasts the
received message with ID (identifier) and its
associated level to sink node.
The sink node sends query by specifying the
attributes (attrs) and how attributes must be
aggregated (agg(expr)) and the sensor nodes that
should be involved in data retrieval is mentioned
using WHERE, GROUPBY, HAVING clauses, and
EPOCH field specifies the period of time (sec) each
sensor should wait before sending its data values.
During the collection phase in which each parent
node has to wait for data to be collected from all
child nodes, Upon receiving all data from child
nodes it aggregated in intermediate node and
transmitted to sink node.
2. DIRECTED DIFFUSION [23]
Direct diffusion is a reactive data centric
approach, which is specifically designed for
applications those requires specific information by
flooding the network with frequent queries. There
are three main phases associated with direct
diffusion interest message, gradient setup, and
path reinforcement and forwarding. Basically,
certain sink nodes propagates interest message
about the information that they are interested in
collecting it from particular sensors (interest
message). Each node upon receiving interest
message, they re-broadcasts it to neighbor nodes.
After that sensor nodes setup interest gradients
that has to be used propagate results back to sink
node (gradient setup). After gradient setup is
complete, there will be only single path for every
source node is reinforced and that path will be
used to forward packets to that path will be used to
forward packets towards the sink node (path
reinforcement and forwarding).
3. EDGE (efficient data gathering protocol)
EDGE (efficient data gathering protocol) is a
data gathering protocol, EDGE is a tree-based
approach rooted towards a sink node. In EDGE
protocol, deletion and addition of nodes requires
the tree reconstruction. EDGE is a
multi-point-to-point approach, in which every
sensor node tries to send data to the sink node. In
EDGE protocol, every node is added to tree
structure by making request and replies. Firstly the
base station broadcasts child requests (CRQ) to all
nodes, then all non-members nodes replies to the
request obtained from base station (CRP). Based on
several metrics associated with non-member node,
best among them will be selected and replied to
accept child (CAC). Then the child will be joined to
tree structure. EDGE protocol is best suited for
applications, where new route found and
reconstruction of tree is required.
4. PEGASIS[24]
Power efficient data gathering in sensor
information system (PEGASIS) is more robust
routing protocol in which sensor nodes are

organized in a chain. Nodes in a chain take turns to
act as a chain leader, where chain leader is one and
only node that is allowed to transmit data directly
to the sink. In this way, it is possible to reduce
communication cost in the network. PEGASIS
proved to be more reliable and efficient, it is also
known as chain based data aggregation algorithm.
The chain construction process requires
knowledge of all nodes in a network, the chain
building process start with node nearer to sink,
then the closest neighbor to this node is selected
and so on.
Chain leader selection takes place according to
following rule: let i be elected leader in round i,
there are N nodes in network {1, 2….N} among
them node will be selected in TDMA schedule. For
this type of scheme a direct communication
channel from each sensor in a network to the sink
is required. In PEGASIS, each node receives data
from its neighbors and aggregates with its own data
and aggregated packet of some length it
transmitted to next neighbor until it reaches to
chain leader. At the chain leader, it includes its
own data into received packet and then aggregated
single packet of data is directly transmitted into the
sink.
5. PEDAP
Power-efficient data gathering and aggregation
protocol (PEDAP), these are data aggregation
scheme based on construction of minimum
spanning tree. PEGASIS and PEDAP two data
aggregation algorithm with same procedure but
PEDAP is the power-aware version of PEGASIS.
Firstly, a node is selected as base-station after that,
the tree starts to build in network of nodes with
minimum weighted edge. The node that wants to
transmit its values will attach to the constructed
tree and transmits its data through the indicated
edge. This process is repeated until all nodes get
attached to tree and transmits there data through
the indicated edge. The constraint that is
associated with PEDAP is the sensors should be in
fixed location. The sensors will sense data
periodically from their associated environment and
aggregate the data and transmits the aggregated
values to selected base-station in round basis. This
protocol can save much energy and show improved
efficiency than PEGASIS.
B. Cluster-Based Approaches
Cluster-based approaches [25] [26] [27] [28] are
similar to tree-based approaches, it is a
hierarchical organization of network in which
nodes are classified into particular cluster on types
of data gathered in them. Some special nodes are
selected has cluster heads(CH) based on some
criteria’s will perform aggregation function on
obtained data from all nodes of particular cluster
and transmits aggregated data into the sink node.
1. LEACH[25]
Low-energy adaptive clustering hierarchy
(LEACH) is a cluster topology data aggregation
algorithm. The important goal of this algorithm is
to reduce the communication cost of data
transmission from normal nodes by having cluster
heads for every particular cluster, where cluster
heads act as aggregator points. There are two main
phase associated with LEACH protocol setup phase
is to form clusters and steady phase is to transmit
data to sink node. Firstly nodes will organize to
form clusters, within each cluster a special node is
elected as cluster head where data is aggregated
which is collected from other nodes in particular
cluster. The cluster head selection is based on
distributed probabilistic approach for current
round of data transmission.
Actual data transmission takes place in second
phase in which all nodes in a cluster send their
sensed data from associated environment to
cluster head. The TDMA protocol is used to avoid
collision among nodes in cluster during data
transmission. After receiving all data from source
nodes, the cluster head performs aggregation on
obtained data and transmits it to sink using single
direct transmission.
LEACH provides the following communication
cost reduction key areas:
 No overhead in selection of cluster heads.
 TDMA protocol is used during data
transmission to avoid collision.
 Minimizing communication cost of each
node is calculated to communicate with its
clusters.
2. DEDA
Delay-minimized energy-efficient data
aggregation, DEDA is a distributed and
energy-efficient data gathering algorithm with
minimum delay. As power-consumption and delay
are two constraints in wireless sensor networks. In
this approach based on data gathered in nodes
they are classified into clusters with particular
cluster (CH), and nodes are cluster members (CM),
and there will be data link between CH and
base-station (BS).
In this approach any two same sized clusters
are joined together to form bigger clusters, and the

process is repeated until clusters cannot be joined
with same sized cluster. Then these joined clusters
form direct connection with base station. Finally
the network consists of clusters of different sizes.
The straight forward scheduling algorithm is
applied and each node is assigned with particular
time-slot to transmit data by its rank, every node of
particular cluster have different ranks with
corresponding different time-slots. By using DEDA
schema minimum delay can be achieved with
reduction in communication cost.
C. Multi-Path Approach
Multi-path approaches are used to overcome
the robustness problem of aggregation trees [29],
[30], [31]. In aggregation trees each node sends
partial result to its single parent, but in these
approaches sends data over multiple paths. The
main strategy of multi-path approaches is that
each node can send data through its multiple
neighbors by using broadcast characteristics of
wireless medium. An aggregation structure used in
multi-path approach is ring topology where nodes
are distinguished into several levels according to
number of hops separating them from sink node.
Data aggregation will be performed on data in
multiple levels towards the sink node. Multi-path
approach helps to transmit duplicates of same
information.
1. SYNOPSIS DIFFUSION[8]
The authors of [8] present the synopsis
diffusion protocol in which data aggregation is
performed through a multi-path approach.
Synopsis diffusion consists of two main phases
such as, distribution of queriesphase and data
retrieval phase. Whenever a nodes send query over
a network leads to formation of ring topology in
synopsis diffusion protocol. There are two different
structures to be considered in synopsis diffusion.
Firstly, a simple ring structure in which during
query distribution phase the network nodes form a
set of rings across a query node Q. another type of
topology has improvements over simple ring
topology, its robust and can cope with changes in
network is called as adaptive topology. In synopsis
diffusion using one of these two topologies, the
data aggregation on obtained data from nodes will
be performed in multiple levels towards the sink
node to reduce communication cost.
V. CONCLUSION
In this survey paper we have presented a detailed
review of data aggregation techniques for
communication cost reduction in wireless sensor
networks. One of the important design issues for
wireless sensor network architectures is energy
efficiency, to keep the network operational as long
as possible to accomplish the requirement of
deployment. Therefore, data aggregation
techniques are an essential building block, as they
aim at reducing the communication cost by
decreasing number of transmissions required for
data collection which, in turn reduces energy
consumption.
REFERENCES
[1] V. Raghunathan, C. Schurghers, S. Park, M.
Srivastava, “Energy-aware Wireless Microsensor
Networks”, IEEE Signal Processing Magazine, March
2002, pp. 40-50.
[2] G. Pottie, W. Kaiser, “Wireless Integrated Network
Sensors, Communication of ACM, Vol. 43, N. 5, pp.
51-58, May 2000.
[3] G. Anastasi, M. Conti, M. Di Francesco, A.
Passarella, “How to Prolong the Lifetime of Wireless
Sensor Networks”, Chapter 6 in Mobile Ad Hoc and
Pervasive Communications, (M. Denko and L. Yang,
Editors), American Scientific Publishers, to appear,
available at
http://info.iet.unipi.it/~anastasi/papers/Yang.pdf
[4] M. C. Vuran, O. B. Akan, and I. F. Akyildiz,
“Spatio-Temporal Correlation: Theory and
Applications for Wireless Sensor Networks”,
Computer Networks Journal, Vol. 45, No. 3, pp.
245-261, June 2004.
[5] J. Li, P. Mohapatra, “Analytical Modeling and
Mitigation Techniques for the Energy Hole Problem
in Sensor Networks”, Pervasive Mobile Computing,
3(3):233-254, June 2007.
[6] D. Ganesan, A. Cerpa, W. Ye, Y. Yu, J. Zhao, D.
Estrin, “Networking Issues in Wireless Sensor
Networks”, Journal of Parallel and Distributed
Computing, Vol. 64 (2004) , pp. 799-814.
[7] A. Mainwaring, J. Polastre, R. Szewczyk, D. Culler
and J. Anderson, “Wireless Sensor Networks for
Habitat Monitoring”, Proc. ACM Workshop on
Wireless Sensor Networks and Applications, pp.
88-97, Atlanta (USA), September 2002.
[8] A. Warrier, S.Park J. Mina and I. Rheea, “How much
energy saving does topology control offer for wireless
sensor networks? – A practical study”, Elsevier/ACM
Computer Communications, Vol. 30 (14-15), Pp.
2867-2879, 15 October 2007.
[9] I. F. Akyildiz and I. H. Kasimoglu, “Wireless Sensor
and Actor Networks: Research Challenges”, Ad Hoc
Networks Journal, Vol. 2, No. 4, pp. 351-367,
October 2004
[10] I. Papadimitriou and L. Georgiadis, “Energy-aware
Routing to Maximize Lifetime in Wireless Sensor
Networks with Mobile Sink”, Journal of

Communications Software and Systems, Vol. 2, No.
2, pp. 141-151, June 2006.
[11] Z. M. Wang, S. Basagni, E. Melachrinoudis, C.
Petrioli, “Exploiting Sink Mobility for Maximizing
Sensor Networks Lifetime”, Proc. of the 38th Annual
Hawaii International Conference on System Sciences
(HICSS '05), Hawaii, 03-06 Jan. 2005.
[12] S. S. Pradhan, K. Ramchandran, “Distributed
source coding using syndromes (DISCUS): design
and construction”,IEEE Transactions on
Information Theory, Vol. 49 (3), Pp. 626-643, March
2003.
[13] C. Tang, C. S. Raghavendra, “Compression
techniques for wireless sensor networks”, Chapter
10 in book WirelessSensor Networks, Pp. 207 - 231,
Kluwer Academic Publishers, 2004.
[14] M. Wu, C. W. Chen, “Multiple Bit Stream Image
Transmission over Wireless Sensor Networks”,
Chapter 13 in book Sensor Network Operations, Pp.
677-687, IEEE & Wiley Interscience, 2006.
[15] Z. Xiong; A. D. Liveris, S. Cheng, “Distributed
source coding for sensor networks”, IEEE Signal
Processing Magazine, Vol. 21 (5), Pp. 80-94, Sept.
2004.
[16] K. Akkaya and M. Younis, “A survey of routing
protocols in wireless sensor networks,” Elsevier Ad
Hoc Network Journal, vol. 3, no. 3, pp. 325–349,
May 2005.
[17] J. N. Al-Karaki and A. E. Kamal, “Routing
techniques in wireless sensor networks: a survey,”
IEEE Wireless Communications, vol. 11, no. 6, pp.
6–28, Dec. 2004.
[18] I. Solis and K. Obraczka, “Isolines: energy-efficient
mapping in sensor networks,” in IEEE ISCC 2005,
Cartagena, Spain, Jun. 2005.
[19]V. Erramilli, I. Matta, and A. Bestavros, “On the
interaction between data aggregation and topology
control in wireless sensor networks,” in IEEE
SECON 2004, Santa Clara, CA, US, Oct. 2004.
[20]H. Luo, J. Luo, Y. Liu, and S. Das, “Energy efficient
routing with adaptive data fusion in sensor
Networks.” in Third ACM/SIGMOBILE Workshop on
Foundations of Mobile Computing, Cologne,
Germany, Aug. 2005.
[21]H. Gupta, V. Navda, S. R. Das, and V. Chowdhary,
“Efficient Gathering of Correlated Data in Sensor
Networks,” in ACM MobiHoc 2005,
Urbana-Champaign, IL, US, May 2005.
[22]S. Madden, M. J. Franklin, J. M. Hellerstein, and W.
Hong, “TAG: a Tiny AGgregation Service for Ad-Hoc
Sensor Networks,” in OSDI 2002, Boston, MA, US,
Dec. 2002.
[23] C. Intanagonwiwat, R. Govindan, D. Estrin, J.
Heidemann, and F. Silva, “Directed diffusion for
wireless sensor networking,” IEEE/ACM Trans.
Networking, vol. 11, no. 1, pp. 2–16, Feb. 2002.
[24] S. Lindsey, C. Raghavendra, and K. M. Sivalingam,
“Data GatherinAlgorithms in Sensor Networks using
Energy Metrics,” IEEE Trans. Parallel Distrib. Syst.,
vol. 13, no. 9, pp. 924–935, Sep. 2002.
[25] W. B. Heinzelman, A. P. Chandrakasan, and H.
Balakrishnan, “An application-specific protocol
architecture for wireless microsensor networks,”
IEEE Trans. Wireless Commun., vol. 1, no. 4, pp.
660–670, Oct. 2002.
[26] Y. Yao and J. Gehrke, “Query processing for sensor
networks,” in ACM CIDR 2003, Asilomar, CA, US,
Jan. 2003.
[27] B. Zhou, L. H. Ngoh, B. S. Lee, and C. P. Fu, “A
Hierarchical Scheme for Data Aggregation in Sensor
Network,” in IEEE ICON 2004, Singapore, Nov.
2004.
[28] A. Mahimkar and T. S. Rappaport, “SecureDAV: A
secure data aggregation and verification protocol for
sensor networks,” in IEEE GLOBECOM 2004,
Dallas, TX, US, Nov. 2004.
[29] S. Nath, P. B. Gibbons, Z. R. Anderson, and S.
Seshan, “SynopsisDiffusion for Robust Aggregation
in Sensor Networks,” in ACM SenSys 2004,
Baltimore, MD, US, Nov. 2004.
[30] A. Manjhi, S. Nath, and P. B. Gibbons, “Tributaries
and Deltas:Efficient and Robust Aggregation in
Sensor Network Stream,” in ACM SIGMOD 2005,
Baltimore, MD, US, Jun. 2005.
[31] S. Chen and Z. Zhang, “Localized algorithm for
aggregate fairness in wireless sensor networks.” in
ACM/SIGMOBILE MobiCom 2006, Los Angeles, CA,
US, Sep. 2006.

Communication Cost Reduction by Data Aggregation: A Survey

More Related Content

What's hot (20)

Viewers also liked (10)

Similar to Communication Cost Reduction by Data Aggregation: A Survey (20)

Recently uploaded (20)

Communication Cost Reduction by Data Aggregation: A Survey