Académique Documents
Professionnel Documents
Culture Documents
Year :2012-2013
IV Year B.Tech (CSE) II Sem
Unit-III PPT Slides
Text Books:
1. Ad Hoc and Sensor Networks Theory and Applications, Carlos Corderio
Dharma P.Aggarwal, World Scientific Publications, March 2006, ISBN 981
- 256-681-3
2. Wireless Sensor Networks: An Information Processing Approach, Feng
Zhao, Leonidas Guibas, Elsevier Science, ISBN 978-1-55860-914-3
( Morgan Kauffman)
Table of Contents
Introduction
The Broadcast Storm
Broadcasting in a MANET
Flooding-Generated Broadcast Storm
Redundancy Analysis
Rebroadcasting Schemes
Multicasting
Issues in Providing Multicast in a MANET
Multicast Routing Protocols
Comparison
Geocasting
Geocast Routing Protocols
Comparison
Conclusion and Future Directions
Routing Protocols
Topology-Based:
Depends on the information about existing links
Position-Based Approaches
Proactive (or table-driven)
Proactive (Table-driven) protocols
Traditional distributed shortest-path protocols
Maintain routes between every host pair at all
times
Based on periodic updates; High routing
overhead
Example: DSDV (destination sequenced distance
vector)
Reactive (On-Demand) protocols
Routing Approaches
Broadcasting
Is a common operation in many applications, e.g., graph-related and
distributed computing problems
It is also widely used to resolve many network layer problems
Multicasting
Transmission of datagrams to a group of hosts identified by a single
destination address
Hence is intended for group-oriented computing
Can efficiently support a variety of applications that are characterized
by close collaborative efforts
Geocasting
Aims at delivering data packets to a group of nodes located in a
specified geographical area
Can be seen as a variant of the conventional multicasting problem, and
distinguishes itself by specifying hosts as group members within a
specified geographical region
Since all these three do communication to a group of recipients, it is
imperative to determine what is the best way to provide these services in
an ad hoc (MANET) environment
Comparison of broadcast, multicast and geocast protocols for ad hoc
networks provided
Example (simple
flooding)
Source node
Rebroadcast node
3
3
5
1
9
8
1st step: S
2nd step: 1, 5, 6, 9
3rd step: 2, 3, 7, 8
4th step: 4
Total: 9 Rebroadcast nodes
9
8
Example broadcast
storm problem in a 13
4
Enode MANET
J
3
G
2
I and J
1 2
2
1
1
D and E
contend at
step 2
A and D collide
at C at step 2
4
C
A
2
2
3
collide at M
at step 4
H and K contend
at step 5
Message still
propagates ..
Redundancy Analysis
S
1
S
1
2
2
2
2
D
D
r
r
Rebroadcasting Schemes
Minimize number of retransmissions of a broadcast
message
Attempt to deliver a broadcast packet to each and
every node in the network
Common Attributes of Broadcast Protocols
Jitter and Random Delay Timer (RDT)
Loop Prevention
Categories of
Broadcasting Protocols
Simple flooding
Probability-based methods
Area-based methods
Neighbor knowledge methods
Categories of
Broadcasting Protocols
Area-based methods
If a sender is located only one meter away, the additional area covered
by the retransmission by a receiving node rebroadcasts is quite low
Distance-based scheme compares the distance between itself and each
of its neighbor nodes that has previously rebroadcast a given packet
Location-based scheme uses a more precise estimation of expected
additional coverage area
Neighbor knowledge methods
Flooding with Self Pruning requires each node to have knowledge of
its one-hop neighbors
Scalable Broadcast Algorithm (SBA) requires that all nodes have
knowledge of their neighbors within a two-hop radius
Each node searches its neighbor tables for the maximum neighbor
degree of any neighbor node, say, dNmax
Calculates a Random Time Delay (RDT)
based on the ratio of
where dme is the number of current
neighbors for the node
d N max
d me
Categories of
Broadcasting Protocols
Neighbor knowledge methods..
Dominant Pruning: Uses two-hop neighbor knowledge, but
proactively choosing some or all of its one-hop neighbors as
rebroadcasting nodes
Multipoint Relaying: Similar to Dominant Pruning, but
rebroadcasting nodes, [Multipoint Relays (MPRs)] are explicitly
chosen by upstream senders
Categories of
Broadcasting
Protocols
Ad Hoc Broadcast
Protocol (AHBP)
be connected as well
Considers the set of higher priority BRGs selected by the previous sender
Dominated Set
Connected Dominating
Set
A connected dominating set (CDS) of
A dominating set of a graph
G= (V, E) is a vertex subset
SV , such that every
vertex vV is either in S or
adjacent to a vertex of S
Broadcast Nodes in
Large Networks?
Some nodes are highly
Categories of
Broadcasting Protocols
Connected Dominating Set-Based Broadcast Algorithm MHs into four groups based on local information as follows:
Group 1: Nodes with degrees larger than the degree of all
neighboring nodes, where connectivity represents the
number of neighbors within the directed transmitting
range of a reference node
Group 2: Nodes have a majority of neighbors with
smaller degree than the reference nodes
Group 3: Remaining nodes not belonging to groups 1, 2
and 4
Group 4: Nodes with degrees smaller than all the
neighbors
A message forwarder is assigned in a decreasing
probability order as p1, p2, p3 and p4
Connected Dominating
Set-Based Broadcast
If A be the area of theAlgorithm
MANET
N be the number of mobile nodes
N 1
P1 pi
i
i 1
N 1
N 1
P3
i 1
i
k i / 2 k
i 1
i 1
1 p j
j 1
j 1
k
i k
N 1
i 1
p j p j
j i
j 1
N 1
k 1 k
P2 pi
i 1
N 1
P4 pi
i 1
i
N 1
j i
i
N 1
j i 1
N 1
i/2
N 1
i k
i 1
p
j 1
1 p j
j i 1
N 1i
Forwarding Probability
Group Forwarding
Probabilities for N=100
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
P1
P2
P3
P4
0.2
0.4
0.6
0.8
1.2
N F N 1 P1 2 P2 3 P3 4 P4
where i is the forwarding probability for the mobile node in group
Lightweight and Efficient Network-Wide Broadcast: Relies on twohop neighbor knowledge obtained from hello packets; however
instead of a node explicitly choosing other nodes to rebroadcast, the
decision is implicit
Lightweight and
Efficient Network Relies on two-hop
Wide
Broadcast
neighbor
knowledge obtained from
hello packets
Each node decides to rebroadcast based on
knowledge of which of its other one and two-hop
neighbors are expected to rebroadcast
Knowledge of which neighbors have received a packet
from the same source node, and which neighbors
have a higher priority for rebroadcasting
The priority is proportional to the number of
neighbors of a given node
The higher a nodes degree is, the higher is its priority
Multicasting v/s
Broadcastng?
Multicasting
Routing protocols offering efficient multicasting in wired
networks may fail to keep up with node movements and
frequent topological changes
Broadcast protocols cannot be used either as multicasting
requires a selected set of nodes to receive the message
All multicast algorithm depend on the topology of the
network
Majority of applications requiring rapid deployment and
dynamic reconfiguration, need multicasting such as
military battlefields, emergency search and rescue sites,
classrooms, and conventions
Crucial to reduce the transmission overhead and power
consumption
Multicasting in a MANET faces challenges such as
dynamic group membership and update of delivery path
due to MH movement
Multicast Routing
Protocols
Broadcast protocols cannot be used as
multicasting requires a selected set of MHs
to receive the message
Protocols are classified into four categories
based on how route to the members of the
group is created:
Tree-Based Approaches
Meshed-Based Approaches
Stateless Multicast
Hybrid Approaches
Source
0
1
2
1
2
2
3
3
3
Tree-Based
Approaches
Extend tree-based
approach to provide multicast in a
MANET
A packet traverses each hop and node in a tree at most once
Very simple routing decisions at each node
Tree structure built representing shortest paths amongst
nodes, and a loop-free data distribution structure
Even a link failure could mean reconfiguration of entire tree
structure, could be a major drawback
Consider either a shared tree or establish a separate tree per
each source
Tree-Based
Ad hoc Multicast Routing Protocol Utilizing Increasing Id-Numbers
Approaches
Tree-Based Approaches
Multicast AODV
Multicast Ad hoc On-Demand Distance Vector
Protocol
Follows directly from the unicast AODV
Multicast
Member
Core
Tree
Mesh
Core-Based
Approaches
Lightweight Adaptive
Multicast (LAM)
Takes into consideration the hierarchical structure used by the ZRP unicast
routing protocol
Network is partitioned into zones
Each node computes its own zone, determined by nodes lieing within a certain
radius of the node
ZRP is described as a hybrid approach between the proactive and reactive routing
protocols
Routing is proactive inside the zones and reactive between the zones
To create a zone, a MZR node A broadcasts an ADVERTISEMENT message with
a time-to-live (TTL) equal to a pre-configured ZONE-RADIUS
Node B within the zone radius decrements the TTL and forwards the message if
appropriate
Node B makes an entry in its routing table for node A, with the last hop of the
ADVERTISEMENT message as the next hop towards destination
Nodes ZONE-RADIUS hops away from node A become border nodes, and serve as
a gateway between node As zone and the rest of the network
MZR begins its search within the zone before extending outward
A source wants to start sending multicast traffic, it initiates the construction of a
multicast tree
A TREE-CREATE-ACK packet is sent back to the source and intermediate nodes
mark in their routing tables the last hop of the TREE-CREATE-ACK as a
downstream node
Multicast Zone
Routing
Source Specific
Multicast Tree
Multicast Optimized Link State Routing (MOLSR)
Other Protocols
The Associativity-Based Ad Hoc Multicast (ABAM)
A spiral fat tree is built as the multicast tree to increase the stability of the tree structure
By using link redundancy of the fat tree, failed links will be easily replaced.
Spiral-fat-tree-based
On-demand Multicast
(SOM)
protocol
A spiral fat tree is built to increase the
stability of the tree structure
Tree
Fat-Tree
Fat-Tree
Mesh based
Approach
multicast
protocols may
Mesh-based
have
multiple paths between any source and receiver
pairs
Mesh-based protocols seem to outperform treebased proposals due to availability of
alternative paths
A mesh has increased data-forwarding
overhead
The redundant forwarding consumes more
bandwidth
The probability of collisions is higher when a
larger number of packets are generated
On-Demand Multicast
Routing
Protocol
Mesh-based protocol employing a forwarding group
concept
Only a subset of nodes forwards the multicast packets
A soft state approach is taken in ODMRP to maintain
multicast group members
No explicit control message is required to leave the
group
The group membership and multicast routes are
established and updated by the source on demand
If no route to the multicast group, a multicast source
broadcasts a Join-Query control packet to the entire
network
This Join-Query packet is periodically broadcasted to
refresh the membership information and updates routes
On-Demand Multicast
Routing Protocol
After
establishing
a
forwarding group and route
construction process, a source
can multicast packets to
receivers via selected routes
and forwarding groups
To leave the group, source
simply stops sending JoinQuery packets
If a receiver no longer wants
to receive from a particular
multicast group, it does not
send the Join-Reply for that
group
Core-Assisted Mesh
multicastingProtocol
by creating a shared mesh
Supports
for
each multicast group
Meshes thus created, helps in maintaining the
connectivity to the multicast users, even in case of node
mobility
It borrows concepts from CBT, but the core nodes are
used for control traffic needed to join multicast groups
Assumes a mapping service by building and maintaining
the multicast mesh
Nodes are classified as: simplex, duplex and non-member
CAMP uses a receiver-initiated method for routers to
join a multicast group
CAMP ensures the mesh to contain all reverse shortest
paths between a source and the recipients
Stateless Approaches
Differential Destination Multicast
Stateless Approaches
DSR Simple
Protocol
Multicast
and
Broadcast
Hybrid Approaches
Ad hoc Multicast Routing Protocol
Hybrid Approaches
Ad hoc Multicast Routing Protocol
Hybrid Approaches
Multicast Core-Extraction Distributed Ad Hoc
Routing
Comparison of Multicast
Approaches
Geocasting
Geocasting is a variant of the conventional multicasting
problem
Group members are within a specified geographical region
Whenever a node in the geocast region receives a geocast
packet, it floods the geocast packet to all its neighbors
A geocast protocol works if at least one node in the geocast
region receives the geocast packet
Protocols use a jitter technique in order to avoid two packets
colliding with each other by a broadcast
Existing geocast protocols divided into two categories: datatransmission oriented protocols and routing creation oriented
protocols
The difference is how they transmit information from a source
to one or more nodes in the geocast region
Data-Transmission
Oriented Geocast
Location-Based
Routing
Multicast Protocols
LBM Scheme 1
A node receives a
geocast packet, it
forwards the packet to
its neighbors if it is
within a forwarding zone
The size of the
forwarding zone
depends on (i) the size of
the geocast region and
(ii) the location of the
sender
A parameter d provides
additional control on
the size of the
forwarding zone
Box forwarding zone: Rectangle covering
source and the forwarding zone
LBM Scheme 2
Does not have a
forwarding zone explicitly
Forwarding is based on
the position of the sender
node and the position of
the geocast region
Node B forwards a geocast
packet from node A if
node B is at least d
closer to the center of the
geocast region
Node K will, however,
discard a geocast packet
transmitted by node B
Voronoi Diagram
The Voronoi diagram
partitions the area in
to a set of convex
polygons such that all
polygon edges are
equidistance
Geo
area
GeoGRID
GeoGRID protocol uses location information in defining
forwarding zone and elects a special host in each grid
area responsible for forwarding geocast packets
The forwarding zone in LBM incurs unnecessary packet
transmissions
A tree-based solution is prohibitive in terms of control
overhead
GeoGRID partitions the geographic area into twodimensional logical grids of size d X d
Two schemes on how to send geocast packets in
GeoGRID: Flooding-Based GeoGRID and Ticket-Based
GeoGRID.
Flooding-Based
GeoGRID
Only gateways in every
grid within the
forwarding zone
rebroadcast the received
geocast packets
A total of m + n tickets
are created by the source
if the geocast region is a
rectangle of m X n grid
Flooding-Based
GeoGRID
Conclusions and
Future Directions
Scalability
Applications for broadcast, multicast,
and geocast over MANETs
QoS
Address configuration
Security
Power control
Table of
Introduction
Contents
TCP Protocol
Overview
Mobility-Related
Fairness-Related
Introduction
TCP most widely used transport protocol
Ad hoc networks composed exclusively of wireless
links
All nodes can move freely and unpredictably
TCP needs to distinguish the nature of errors
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
67
68
TCP Basics
Byte Stream Delivery: TCP interfaces between the application
layer above and the network layer below and TCP decides
whether to segment or delineate the byte stream in order to
transmit data in manageable pieces to the receiver, hence called
byte stream delivery service
Connection-Oriented: Two communicating TCP entities (the
sender and the receiver) must first agree upon the willingness to
communicate
Full-Duplex: TCP almost always operates in full-duplex mode,
and TCP exhibit asymmetric behavior only during connection
start and close sequences (i.e., data transfer in the forward
direction but not in the reverse, or vice versa)
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
69
Timers: TCP maintains various static and dynamic timers on data sent
and if the timer expires before receiving an acknowledgement, the
sender can retransmit the segment
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
70
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
71
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
72
Control Bits
Urgent Pointer (URG) If this bit field is set, the receiving TCP should
interpret the urgent pointer field
Acknowledgement (ACK) If this bit is set, the acknowledgment field is valid
Push Function (PSH) If this bit is set, the receiver should deliver this segment
to the receiving application as soon as possible
Reset Connection (RST) If this bit is present, it signals the receiver that the
sender is aborting the connection and all the associated queued data and
allocated buffers can be freely relinquished
Synchronize (SYN) When present, this bit field signifies that the sender is
attempting to synchronize sequence numbers
No More Data from Sender (FIN) If set, this bit field tells the receiver that the
sender has reached the end of its byte stream for the current TCP connection
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
73
TCP Details
Window: This is a 16-bit integer used by TCP for flow control in the form
of a data transmission window size
Checksum: A sender computes the checksum value of 16-bits, based on the
contents of the TCP header and data fields and is compared with the value
the receiver generates using the same computation
Urgent Pointer: This 16-bit field tells the receiver when the last byte of
urgent data in the segment ends
Options: Depending on the option(s) used, the length of this field varies in
size, but it cannot be larger than 40 bytes due to the maximum size of the
header length field (4 bits)
Padding:It may be necessary to pad the TCP header with zeroes so that
the segment ends on a 32-bit word boundary as defined by the standard
Data: This variable length field carries the application data from TCP
sender to receiver and this field coupled with the TCP header fields
constitutes a TCP segment
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
74
Congestion Control
Slow Start: (A mechanism to control the transmission rate) Whenever a
TCP connection starts, the Slow Start algorithm at the sender initializes a
congestion window (CWND) to one segment and the congestion window
increases by one segment for each acknowledgement returned
Congestion Avoidance: When Slow Start forces a network to drop one or
more packets due to overload or congestion, Congestion Avoidance is used
to reduce the transmission rate
Fast Retransmit: When a duplicate ACK is received, the sender does not
know if this is because a TCP segment was lost or because a segment was
delayed and received out of order at the receiver and if more than two
duplicate ACKs are received by the sender, it does not even wait for the
Retransmission Timeout to expire and retransmits the segment (as
indicated by the position of the duplicate ACK in the byte stream)
Fast Recovery: The sender has implicit knowledge that there is data still
flowing to the receiver since duplicate ACKs can only be generated when a
segment is received and the sender only enters Congestion Avoidance mode
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
75
Time Estimation
Round-Trip Time Estimation: When a host transmits a TCP packet to
its peer, and the reply does not come within the expected period, the
packet is assumed to have been lost and the data is retransmitted
Over an Ethernet, no more than a few microseconds should be needed
for a reply
This process called Round-Trip Time (RTT) estimation
If the RTT estimate is too low, packets are retransmitted
unnecessarily; if too high, the connection can sit idle while the host
waits to timeout
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
76
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
77
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
78
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
79
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
80
Reestablishing Path
5
6
4
The routing protocol reestablishes the path through node 6
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
81
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
82
Each of these frames carries the remaining duration of time for the
transmission completion, so that other nodes in the vicinity can hear it
and postpone their transmissions
The nodes must await an IFS interval and then contend for the medium
again
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
83
Consider a linear topology in which each node can only communicate with its
adjacent neighbors
In addition, consider that in Figures 7.7(a) and 7.7(b) there exist a single TCP
connection running between nodes 1 and 5
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
84
TCP Throughput
Larger the number of nodes a TCP connection needs to span, lower is
the end-to-end throughput, as there will be more medium contention
taking place in several regions of the network
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
85
Capture Conditions
86
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
87
DSR
DSR protocol operates on an on-demand basis in which a node
wishing to find a new route broadcasts a RREQ packet
The problem with this approach concerns the high probability of
stale routes in environments where high mobility as well as
medium constraints may be normally present
The problem is exacerbated by the fact that other nodes can
overhear the invalid route reply and populate their buffers with
stale route information
It can be mitigated by either manipulating TCP to tolerate such
a delay or by making the delay shorter so that the TCP can deal
with them smoothly
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
88
TORA
TORA has been designed to be highly dynamic by establishing
routes quickly and concentrating control messages within a small set
of nodes close to the place where the topological change has
occurred
TORA makes use of directed acyclic graphs, where every node has a
path to a given destination and established initially
This protocol can also suffer from stale route problem similar to the
DSR protocol
The problem occurs mainly because TORA does not prioritize
shorter paths, which can yield considerable amount of out-ofsequence packets for the TCP receiver, triggering retransmission of
packets
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
89
90
Route Assymetry
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
91
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
92
93
Fixed Retransmission
Timeout (RTO)
Relies on the idea that routing error recovery should be
accomplished in a fast fashion by the routing algorithm
It disables such a mechanism whenever two successive
retransmissions due to timeout occur, assuming that it
actually indicates route failure
TCP sender doubles the RTO once and if the missing packet
does not arrive before the second RTO expires, the packet is
retransmitted again and again, but the RTO is no longer
increased
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
94
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
95
TCP-DOOR
(Detection Out-OfOrder and Response )
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
96
Main Drawbacks
The approaches that rely on feedback information from inside the
network (TCP-F, ELFN-based, ATCP) may fail in situations where
TCP sender is unable to receive data from the next hop node
The usage of explicit notification by the intermediate nodes, such
as ECN, raises many security concerns
The assumption in TCP-DOOR that OOO packets are exclusive
results of route disturbance may not be true in a quite a few
scenarios
The main concern addressed by the approaches presented so far is
how to avoid the TCP exponential backoff mechanism when losses
take place by factors other than congestion
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
97
FairnessRelated
COPAS
A protocol called COPAS (COntention-based PAth
Selection) has been proposed to address TCP performance
drop due to the capture problem and resulting unfairness
COPAS implements two novel routing techniques in order
to contention-balance the network, namely, the use of
disjoint forward (for TCP data) and reverse (for TCP
ACK) paths to reduce the conflicts between TCP packets
traveling in opposite directions
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
98
Route Establishment in
COPAS
TCP ACK
B
2
8
5
7
D
S
3
1
I
TCP Data
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
99
Average Aggregate
Throughput
50 Nodes
100 Nodes
100
Neighborhood RED
(Randomly Early
Two unique
features of ad hoc wireless networks are the key to
Detection)
understand unfair TCP behaviors: Spatial reuse constraint and the
location dependency
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
101
Node As
Neighborhood
and Distributed
Queue
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
102
Copyright 2006, Dr. Carlos Cordeiro and Prof. Dharma P. Agrawal, All rights reserved.
103