A Cross-Layer-Based Routing With Qos-Aware Scheduling For Wireless Sensor Networks

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

A Cross-Layer-Based Routing with QoS-Aware Scheduling for

Wireless Sensor Networks


Hind Alwan and Anjali Agarwal
Department of Electrical and Computer Engineering, Concordia University
Montreal, Quebec, Canada
Email: {h_alwan, aagarwal}@ece.concordia.ca
AbstractDue to the increased use of sensor nodes in a
variety of application fields, wireless sensor networks need to
handle heterogeneous traffic with diverse priorities to achieve
the required quality of service. In this paper, we address the
cross layer quality of service-aware scheduling for wireless
sensor network with respect to delay and reliability in an
energy efficient way. A node disjoint multipath routing is used
and a QoS-aware priority scheduling considering MAC layer is
proposed to ensure that real time and non-real time traffic
achieve their desired QoS while alleviating congestion in the
network. We evaluate our algorithm with extensive simulations
and the results have demonstrated the effectiveness of our
proposed scheme for different metrics.
Keywords- Quality of service routing; scheduling; multipath;
congestion avoidance.

I.

INTRODUCTION

The main challenge involved with routing protocols in


Wireless Sensor Networks (WSNs) is to provide the
required Quality of Service (QoS) [1] while respecting the
resources limitation of the network such as severe resource
constraints of sensor nodes, high node density, unreliable
links and harsh environmental conditions. There are many
other challenges involved with routing protocols for WSNs
since different applications call for different QoS such as
end-to-end delay and reliable data transmission guarantees.
Thus, the design requirements of WSNs change according to
the application [2]. Accordingly, it is difficult to propose a
routing protocol that satisfies the wide range of applications
while respecting the WSNs constraints. Furthermore, the
inter-dependency and conflict among multiple QoS
parameters make the problem difficult and NP-complete.
This calls for a suitable routing protocol tailored to
achieve the application-specific QoS and respect the
characteristics of WSNs. Moreover, an efficient allocation
of network resources to satisfy the different QoS
requirements is the primary aim of a QoS-based routing
protocol.
Multiple applications running on WSNs require the
network to handle traffic with different priority levels and
QoS requirement in an energy efficient way while avoiding
collisions and interference. However, in order to provide the
required QoS in WSNs while considering the unique
properties of sensor networks, energy awareness and robust
protocol design at all layers of the networking protocol stack
[3] is required.

978-1-4799-0792-2/13/$31.00 2013 IEEE

In network layer, the main functions are to provide endto-end data routing and congestion control. Therefore, the
end-to-end requirements guarantee cannot be only provided
by QoS routing in a network layer; it is needed to
investigate the other layers that allocate resources like
Medium Access Control (MAC), which plays a key role in
determining the channel access delay, utilization and energy
consumption. MAC layer coordinates the sharing of the
wireless medium layer and can contribute to energy
efficiency by minimizing the number of collisions,
overhearing, overhead and ideal listening. Therefore, the
MAC layer dominates the performance of the QoS support
in the network [4].
Our proposed scheme extends the routing approach in
[5] and considers the joint functionalities among the layers
especially the routing and MAC layers. A cross-layer design
is proposed between the routing and MAC layers where the
end-to-end QoS requirements are enforced through sensors
decision of next hops according to the neighbors state and
the required QoS. However, the end-to-end requirement is
guaranteed jointly by the local decisions of these sensors
and the sink decision on the used paths and the number of
these paths. The proposed scheme prioritized traffic
according to the requirements such that the end-to-end
requirements can be improved with packet, path and queue
scheduling. When the traffic load on sensors in some area of
the network is high due to heavy communication activity,
the probability of routing through this area is decreased to
protect the traffic from dropping.
The rest of the paper is organized as the following:
Section 2 introduces an overview of existing related works.
Section 3 provides the proposed prioritized scheduling.
Section 4 describes the node-disjoint multipath routing
protocol in details. Section 5 presents the performance
evaluation. Finally, Section 6 concludes the paper.
II.

RELATED WORK

Sensors deployed in WSN are energy limited devices


and therefore energy efficient communication techniques
are the most important requirements in these networks.
Cross-layer design with routing and MAC as two important
candidate layers has been proposed as a solution for
resource constrained WSNs and many researches have been
conducted on this perspective [6-10].
Congestion can degrade the network performance and
obstruct the application requirements. It can cause packet

losses, increased delay, and increased energy consumption.


For example, a node may have many packets backlogged
due to heavy load, and if it is chosen to forward other
packets, it increases the packet latency and may even drop
packets due to queue overflow, which in turn reduces the
higher layers throughput. Accordingly, the timeliness
problem in WSNs is studied from the congestion point of
view. Hence, many solutions have been proposed in the
literature to control the congestion in WSN such as rate
control, queue management, and traffic prioritization.
A real time communication protocol for large-scale
WSNs is presented in [6]. A velocity monotonic scheduling
is introduced that inherently accounts for both time and
distance constraints in order to reduce the end-to-end
deadline miss ratio in sensor network. The velocity of a
packet is calculated based on the end to end deadlines and
the communication distance and assigned priority
accordingly. However, the main drawback of this protocol is
that in the next hop selection process, only greedy
geographic forwarding is considered while the conditions of
the local wireless links are not considered. Therefore, load
balancing and congestion avoidance in packet transmission
are not achieved.
A cluster based QoS aware routing protocol for WSNs is
proposed in [7]. The protocol finds the least cost and energy
efficient path that meets the end-to-end delay. A cost
function is associated with each link considering the link
delay and a class based queuing model is employed to
handle both real time and non-real time traffic. The
bandwidth is shared for real time and non-real time traffic
and is adjusted in order to satisfy the delay requirements.
However, packet collision or loss is not considered in the
design of this protocol.
A node priority based control mechanism for wireless
sensor networks is proposed in [8]. Node priority index is
presented to reflect the importance of each node. The packet
inter arrival time and service time are used to measure the
congestion degree at a node. Moreover, a hop-by-hop
congestion control is used in order to control congestion
faster and in an energy efficient way. However, the protocol
does not consider the sensed data within a node. Moreover,
it does not consider any mechanism to handle prioritized
heterogeneous traffic in the network.
In [9], the proposed protocol provides QoS
differentiation in timeliness and reliability domains for
WSNs based on a cross-layer approach between the network
and the MAC layers. To avoid congestion and decrease the
packet loss rate, packets are transmitted with respect to the
required end-to-end delay parameter. By using the distance
to the sink and delay information, each node calculates the
required speed and selects the next hop such that the speed
requirement is met. And to support reliability, multipath is
used and the number of these paths is based on the required
end-to-end reaching probability.
In [10] a node-disjoint multipath routing protocol is
proposed to provide reliability and delay requirements of

real time applications. The energy, remaining buffer size


and signal-to-noise ratio are used as parameters in the link
cost function to select the next hop through the paths
construction phase. To improve reliability, Forward Error
Correction (FEC) mechanism is used to introduce data
redundancy for data transmission. To achieve the delay
requirements of various applications, a queuing model is
adopted to manage the real time and non-real time traffic.
In our work, we combine different ideas from the
previously proposed work in order to optimally tackle the
problem of QoS in WSNs. In our proposal we try to provide
the required QoS in terms of end-to-end delay and reliability
in an energy efficient way considering the joint
functionalities among the routing and MAC layers.
III.

PROPOSED PRIORITIZED SCHEDULING

MAC layer is responsible for scheduling and allocation


of the shared wireless channel, which eventually determines
the link level QoS parameters specifically MAC delay. This
section presents the QoS-aware scheduling parameters that
reflect the performance of the routing protocol in terms of
end-to-end delay and reliability of data transmission with
minimum resource consumption.
A. Network Model and Definitions
Let the network graph be G = (S, L) where S is the set of
nodes and L is the set of all possible communication links.
Let be node i in S and the link between to where i
and j N and N is the number of nodes in the network. The
following definitions are used in this paper:
is the set of neighbors of , all nodes that are

inside the transmission range, a, of node .

is the expected progress from the sender


,
node and the neighbor node towards the sink,

.
=
, where
,
,
,
and
are the distance from node
,
,
and to the sink respectively. We assume that each
node knows its position and that of the base station.
The available energy at a node is specified as
=
, where
is
and is calculated as
is the energy
the initial energy at the node and
= (
+
consumed per transmission.

) b,
and
are the energy

consumption to transmit and receive one bit of data,
is the energy consumption of
respectively,
power amplifier and b is the number of bits in the
transmitted packet. Therefore, the energy consumed
to transmit one data packet on a path of hop number
=
.
of nodes can be written as;
Queue length is one of the parameter used to estimate
congestion at a node and congestion of a node is
represented as the load on that node. Therefore, we
used link load,
, as one of the node metric as;

(1)

and
are the length of occupied
where
.
.
and maximum buffer of node , respectively.
at a node, the more chance to
The smaller
accept new traffic.
Link reliability, r, is calculated as the signal to noise
ratio (SNR) and is used to choose the node that
achieves high probability of data delivery.

priority queue may not get any service until the highest
priority traffic is served completely, which is commonly
known as the starvation problem.
The available bandwidth, BW, at a wireless link is shared
among these two queues using the weighted round robin
fashion. If the queue has packets to transmit during that time

B. Traffic Classification and Prioritization


To support applications with diverse QoS requirements,
we classify these requirements into four different classes
concerning both delay and reliability. Packets are
, by reading the packet header which
prioritized,
includes a priority number for each type of packet as
follows:
Class 1: for delay sensitive requirements where
packets delivery requires delay constraints only.
= 2.
Class 2: packets delivery requires delay-bound and
= 2.
reliability.
Class 3: this class belongs to normal applications;
packets deliveries require no reliability and delay
=1.
constraints.
Class 4: applications with reliability requirements
= 1.
only.
The proposed traffic prioritization scheme assigns
priorities to traffic at the source node according to the delay
requirement, in order to guarantee the requested end-to-end
delay in multi-hop wireless networks. On each node, there is
a classifier, which checks the type of the incoming packet
and sends it to the appropriate queue as shown in Fig. 1. The
packets that are related to low priority queue, Q1, are the
= 1. The packets that are
non-real time traffic,
related to the high priority queue, Q2, are the delay sensitive
= 2. The length of
packets, the real time traffic,
occupied queue at a node is given as;
=
+
,
where
and
are the length of occupied queue of Q1
and Q2, respectively.
C. Queuing Model
In order to support service differentiation and to provide
QoS for high priority traffic, we use a priority queuing
protocol which prioritizes the packet transmission process at
each node. Queue is used for storing the data temporarily
and the length of queue is one of the parameter which is
used to get an estimate of congestion at the nodes.
Using priority queue, the scheduler of the sensor node is
serving different output queues; the high priority queue is
served first. If there is no packet waiting in the high priority
queue, it will serve the low priority queue. However, if the
amount of highest priority traffic is extreme then the lower

Figure 1. Queue model at a node

slot it transmits the packets. Otherwise, it passes it to the


next queue. Depending on the priority of the queue, we can
calculate the bandwidth assigned for each queue, with
weight , as:
where Q is Q1 or Q2.
IV.

BW

(2)

NODE-DISJOINT MULTIPATH ROUTING

In this section, we present the parameters used in our


proposed link cost function and their influence on providing
the required QoS. Also, we review the node-disjoint
multipath process used and the criteria used to select these
paths.
A. Neighbors Discovery and Next Node Selection
In the neighbor discovery phase, sensor nodes introduce
themselves to their one-hop neighbor nodes by sending

HELLO messages, Fig. 2(a). When a node receives a Hello


message, the node records the received information to
update its neighbor table entries and the information related
to each neighbor in the neighbors set, .

set of the current node ,


is
neighbor ,
the load at node , and , is the SNR on link . The
weight , and are the weights that indicate the
importance of each parameter in selecting the next hop and
( + + ) =1.
B. Node-Disjoint Multipath Selection
In path discovery phase and in order to construct multi
node-disjoint path to the sink, route request message,
RREQ, Fig. 2(b), is initiated at the source node;
1. hop =0; hop is the hop count at the path,
=0;
represents the number of
2.
loaded nodes along the path,
= 0;
is the end-to-end path delay,
3.
4.
= 1;
is the end-to-end path quality.

Figure 2. Control messages structure. (a) Hello message (b) RREQ


message (c) RREP message

Our QoS routing goal is to achieve the requirements in


terms of end-to-end delay and reliability of data
transmission while extend the network life time. To achieve
this goal, the parameters that influence delay, reliability of
data transmission and energy consumption at each hop on
the routing path/paths should be considered. In our proposed
solution, we consider the following parameters in selection
of the next hop;
1. High geographic progress toward the sink. Due to the
profits from geographic routing, we considered the
idea of greedy forwarding in order to minimize the
number of sensor nodes used to route data between
source and destination.
2. High available energy to balance the energy
consumption among the candidate nodes in order to
extend the network lifetime.
3. High link reliability. Link reliability degradation at a
node reflects the interference degree around that node
and can lead to packet losses, which affect the
reliability of data delivery to the sink.
4. Less node congestion. Congestion at a node can lead
to packet losses, increase transmission delay and
influences the energy efficiency.
is computed for each
To decide the next hop,
set and the one with the maximum
neighbor in the
value is selected as the next hop.
=
where

+ /

(3)

is the available energy at the candidate

The source node also reports the application


, and end to
requirements in terms of end-to-end delay,
end data delivery reliability,
, in the RREQ message.
The RREQ is then send to all the neighboring nodes in the
set of the source node.
After receiving the RREQ message, each node in
updates the information of the RREQ as follows
before sending it to the selected neighbor.
1. hop = hop +1,
>
) then
2. If (
+ 1. Otherwise, no change to
.
is the value for load threshold,
=
+ ,
3.
4.
=
.

The RREQ message is then sent to the candidate


value in the
set.
neighbor with the maximum
C. End-to-End QoS Scheduling-Based Routing
By receiving the RREQ messages, the sink estimates the
number of all available node-disjoint paths to the source and
uses the parameters of each path, the maximum available
load of a node at a path, end-to-end delay and path quality to
, as follows;
calculate the cost function of each path,
= (

(4)

is the energy threshold value at a node to


where
<
it
participate in a transaction. When a node has
cannot participate in data routing, thus it is considered dead.
) is used to reduce the
In (4) the load function (1/
cost of a path that suffers from high load.
The sink then evaluates the optimal paths for each traffic
demand as follows:

1. Assign the values of , and according to the


requested requirements.
for all the available paths, p.
2. Calculate
3. Sort available paths according to their cost such that
>
> >
4. A scheduler is used to determine which path to select
for current traffic demands based on the requested
services class. In order to reduce network congestion
and enhance the network performance, classes with
higher priority will be transmitted in routing
. Route reply,
path/paths with the smallest
RREP, message Fig. 2(c) is then sent to the source
node through the selected path/paths. The number of
these paths is decided according to the required
reliability as described next.
D. Number of Routing Paths
In order to guarantee the required end-to-end reliability
level, FEC coding at the source node is used. Thus, by
increasing the number of forwarding paths depending on the
required reliability, the probability of packet received by the
sink grows.
For the classes with low or no reliability requirements,
packet is delivered using a single path. While for the classes
with high reliability requirements, packet is delivered over
, received by a RREQ
multiple paths. The path quality,
messages is used to calculate the end-to-end packet
, as follows,
transmission reliability,
=1-

(5)

The sink uses the packet reliability to determine the


number of multipath, np, and the priority of each path. Each
RREP message carries the priority number of a path for
each traffic demands. The sink transmit RREQ message to
the source node through the selected paths.
From the number of the received RREP messages, the
source node obtains the number of paths to be used and the
priority value of each path. Source node then starts the FEC
coding and fragments are assigned to each path such that the
first fragment is assigned to the path with the highest
priority. The second fragment to the second highest priority
path and so on till the requested reliability is achieved using
Algorithm 1.
V.

PERFORMANCE EVALUATION

In this section, we present the main results from an


extensive performance evaluation of our proposed scheme.
A. Simulation Model and Performance Metrics
We evaluate the performance and validate the
effectiveness of our proposed scheme through simulation.
We adapt the same C++ codes used in our previously
published works [5, 11]. These papers illustrate the validity

Algorithm I. Calculating number of multipath

np = 1;
//Initialization
<
)
if (
{

=
;
for (i = 2; i ; i++)
{
np = np++;

= 1
1
)
if 1

)
{
number of paths to be used = np;
break;
}
}
}

and comparability of our implementation, in which the


validation tests cover the basic functionality of the multipath
QoS-aware routing protocol in WSNs. The evaluations
demonstrate that MQoSR protocol [5] outperforms the
protocols presented in [12, 13]. Therefore, to show the
effectiveness of our queue scheduling mechanism; the
proposed protocol is executed on and compared with the
QoS-aware routing protocol, MQoSR, that does not adapt
scheduling mechanism, using the simulation parameters
presented in Table 1.
A wireless sensor network which comprises of 300 static
sensor nodes is randomly distributed in 200m 200m area.
All sensor nodes have the same transmission radius of 40m.
IEEE 802.11 is used for the MAC layer. It has been widely
adopted and used in both traditional wireless networks and
in multi-hop wireless sensor networks research. Source
nodes are located in the left lower corner and sink node is
located in the right upper corner of the simulation area like
the model shown in Fig. 3. Two sources targeting to a single
sink is considered to generate traffic with a packet size of
512 bytes at CBR (constant bit rate). First source generates
real time traffic, RT, while the second source generates the
non-real time traffic, NRT. We change the total packets
arrival rate at the sources from 2 to 40 packets/s for each
class. Simulation results are obtained from different
configurations (10 runs) to reduce the effect of the position
of sensors. At each point, the results shown are averaged
over 10 simulation runs for all the traffic 20 to 400 packets
for each class (that is 10 runs with 2 to 40 packets each)
with a 90% confidence interval, which is not plotted for the
sake of legibility.
We evaluate mainly the performance according to the
following metrics:
1. On-time packet delivery ratio: The ratio of the
number of packets received at the sink with end-toend delay equal to or less than the required end-to-

packets are sent, real time traffic are given the highest
priority and processed first and this introduces more
queuing delay for non-real-time traffic at each sensor node.
Average End-to-End Delay (second)

0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
10

Figure 3. Simulation model

3.
4.

TABLE I. Simulation Parameters


Network field

200m 200m

Number of sensors

300

Simulation time

100s

MAC layer

IEEE 802.11

Transmission range

40m

Packet size (data + overhead)

128 byte

No. of source nodes

No. of sink

Each queue size

0.9
0.8
0.7
0.6
0.5

0.3
10

Class 1 - RT
20
30

Class 2 - RT
Class 3 - NRT
Class 4 - NRT
40
50
60
70
80
90
100
Packet Arrival Rate (packet/second)

Figure 5. On time reachability

50 packets
50 nJ/bit
50 nJ/bit
100 pJ/(bit.

0.4

2J

30
40
50
60
70
Packet Arrival Rate (packet/second)

Figure 4. Average end-to-end delay

O n-Time Reachability

2.

end delay, to the total number of packets sent by the


source node.
Average packet delivery ratio: The ratio of the
number of packets received successfully at the sink to
the total number of packets transmitted by the
sources.
Average energy consumption: The average energy
consumed to transmit a data packet to the sink.
On time reachability: The probability that a packet
meets the required deadline.

20

Class 1 - RT
Class 2 - RT
Class 3 - NRT
Class 4 - NRT
80
90
100

B. Simulation Results
Figs. 4 and 5 illustrate the average end-to-end delay per
packet and the on time reachability for both real time and
non-real time traffic, respectively. In order to focus on the
timeliness domain, we use a non-strict reliability
requirement of 0.7. Conversely, we use a strict real time
requirement of 50ms. From the results, it is clear that the
average delay increases as traffic rate increases. When more

Fig. 6 shows that the average energy for a sensor node


increases with the packet rate. From the results, we can see
that Class 3 traffic has the least energy consumption among
the other classes even when the arrival rate is increased. In
Class 3 the forwarding strategy used considered the energy
as the main metric as well as the load avoidance technique
adapted to guarantee a fair service to real time and non-real
time traffic. However, it is worth emphasizing that the price
to meet the required QoS is the overhead introduced in
terms of energy consumption.
Fig. 7 shows the performance of the proposed scheme in
terms of the average packet delivery ratio for a strict
reliability requirement of 0.9. We observe that even for
higher loads, most of generated packets achieve their
reliability requirement. More packets are delivered even
under heavy load and this is because the FEC technique is
used to enhance the probability that packets are recovered at
the sink as well as the forwarding strategy that considered

Average Energy Consumption (mJ/node)

load at sensor nodes is employed to alleviate congestion in


the network and ensure that the real time traffic is not only
reported fast but also not lost due to queue overflow at
sensor nodes.
2

1.5

0.5

10

20

30

40
50
60
70
Packet Arrival Rate (packet/second)

Class 1 - RT
Class 2 - RT
Class 3 - NRT
Class 4 - NRT
80
90
100

The results in Fig. 8 show the average end-to-end delay


per packet for each class by each protocol. With the increase
in packet drop probability, MQoSR reaches high end-to-end
delay compared to our scheme and this confirms the
effectiveness of the congestion avoidance strategy adapted
and the priority mechanism used in order to meet the
timeline requirement. Note that the end-to-end delay for
Class 1 and Class 2 are not affected much compared to Fig.
4. In Class 1 packet has the highest priority among other
classes and thus other low priority packet may be dropped
due to congestion. However, In Class 2, multipath routing
with FEC technique is used to deliver packet considering
link quality as well as delay as metrics, thus packet can be
still recovered at the sink when some its sub-packet are
dropped.
Fig. 9 shows the results for the average energy
consumption. Obviously our proposed scheme outperforms
the MQoSR protocol in term of energy consumption. This is
also confirming the energy efficient scheduling mechanism
adapted to achieve the required QoS.

Figure 6. Average energy consumption


Average End-to-End Delay (second)

0.25

Packet Delivery Ratio

0.9
0.8
0.7
0.6

0.2

0.15

0.1

0.05

0.5

0.01

Class 1 - RT
Class 2 - RT
Class 3 - NRT
Class 4 - NRT
20
30
40
50
60
70
80
90
100
Packet Arrival Rate (packet/second)

Figure 7. Packet delivery ratio

In Fig. 8 and Fig. 9, we study the effect of packet drop


probability on the performance of our proposed scheme
compared with MQoSR [5]. The probability of packet drop
is varied from 0.01 to 0.05 and the packets arrival rate is set
to 50 packets/second. In both figures, we use Delay
(MQoSR)
to refer to the applications with delay
requirements used in MQoSR protocol and is equivalent to
Class 2 in this paper and Delay + Reliability (MQoSR) to
refer to the applications with delay and reliability
requirements used in MQoSR protocol and is equivalent to
our Class 2 requirements. Thus, we are comparing the same
routing strategies used for both the classes, nevertheless
MQoSR does not consider congestion avoidance and
prioritised packet scheduling compared to our scheme.

0.02

0.03
Packet Drop Probability

0.04

0.05

Figure 8. Average end-to-end delay


Average Energy Consumption (mJ/packet)

0.4
10

Class 1 - RT
Class 2 - RT
Delay (MQoSR)
Delay + Reliability (MQoSR)

3
2.5
2
1.5
1
0.5

0.01

0.02

Class 1 - RT
Class 2 - RT
Delay (MQoSR)
Delay + Reliability (MQoSR)
0.03
0.04
0.05
Packet Drop Probability

Figure 9. Average end-to-end delay

VI.

CONCLUSION

In this paper, the joint functionalities among the routing


and MAC layers are considered to design a cross-layer
scheme in order to deliver data according to the required
end-to-end QoS. The QoS requirements are enforced
through sensors decision of next hops according to the
neighbors state. However, the end-to-end requirement is
guaranteed jointly by the local decisions of these sensors
and the sink decision on the used paths and the number of
these paths. The proposed scheme prioritized traffics
according to the requirements into a packet, queue and path
scheduling. Moreover, the queue size of each sensor is used
as an indicator of node congestion, and presented in the link
cost function as a metric. In this way the node with the high
load has a lower chance to be selected as next hop. Also, by
transferring this information to the sink and when the load
of traffics on sensors in some area of the network is high
due to heavy communication activity, the cost of routing is
decreased through this area to protect the traffic from
dropping.
References

[4]

M. Aykut Yigitel, O. Durmaz Incel, and C. Ersoy, QoS-Aware MAC


Protocols for Wireless Sensor Networks: A survey, Computer
Networks vol. 55, no. 8, 2011, pp. 1982-2004.

[5] H. Alwan and A. Agarwal, Multi-Objective QoS Routing for


Wireless Sensor Networks, Proc. International Conference on
Computing, Networking and Communications, San Diego, January
2013, pp. 1074-1079.
[6]

C. Lu, B. M. Blum, T. F. Abdelzaher, J. A. Stankovic and T. He,


RAP: A Real-Time Communication Architecture for Large-Scale
Wireless Sensor Networks, Proc. IEEE Real-Time Technology and
Application Symposium, San Jose,CA ,USA, September 2002.

[7]

K. Akkaya and M. Younis, An Energy Aware QoS Routing Protocol


for Wireless Sensor Networks, Proc. International Conference on
Distributed Computing Systems Workshops, Providence, RI, USA,
May, 2003, pp. 710-715.

[8]

C. Wang, B. Li, K. Sohraby, M. Daneshmand and Y. Hu, PCCP:


Upstream Congestion Control in Wireless Sensor Networks Through
Cross-Layer Optimization, IEEE Journal on Selected Areas in
Communications vol. 25, no. 4, May 2007, pp. 786-795.

[9]

E. Felemban, C.G. Lee, and E. Ekici, MMSPEED: Multipath


Multispeed Protocol for QoS Guarantee of Reliability and Timeliness
in Wireless Sensor Networks, IEEE Trans. Mobile Comput., vol. 5,
no. 6, Jun. 2006, pp. 738-754.

[10] J. B. Othman and B. Yahya, Energy Efficient and QoS Based


Routing Protocol for Wireless Sensor Networks, J. Parallel Distrib.
Comput., vol. 70, no. 8, 2010, pp. 849-857.

[1]

J. Yick, B. Mukherjee, and D. Ghosal, Wireless Sensor Network


Survey, Computer Networks, vol. 52, 2008, pp. 22922330.

[11] H. Alwan and A. Agarwal, Multi-objective Reliable Multipath


Routing for Wireless Sensor Networks, Proc. IEEE Globecom
Workshop on Ad Hoc and Sensor Networking, Florida, December
2010, pp. 1227-1231.

[2]

F. Xia, QoS Challenges and Opportunities in Wireless


Sensor/Actuator Networks, Sensors, vol. 8, no. 2, 2008, pp. 10991110.

[12] X. Huang and Y. Fang, Multi-constrained QoS Multipath Routing in


Wireless Sensor Networks, ACM Wireless Networks, vol. 14, 2008,
pp. 465-478.

[3]

B. Yahya and J. Ben-Othman, Towards a Classification of Energy


Aware MAC Protocols for Wireless Sensor Networks, Journal of
Wireless Communications and Mobile Computing, vol. 9, no. 12,
2009, pp. 1572-1607.

[13] A. B. Bagula and K.G. Mazandu, Energy Constrained Multipath


Routing in Wireless Sensor Networks, Proc. of Conference on
Ubiquitous Intelligence and Computing, vol. 5061, 2008, pp. 453467.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy