Ly3421472152

Mrs. S. Jeevitha / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 3, Issue 4, Jul-Aug 2013, pp.2147-2152
2147 | P a g e
Variable Workload Scheduling With Increased Throughput in
Wireless Sensor Networks
Mrs. S. Jeevitha
M.E Software Engineering, Mount Zion College of Engineering and Technology, Pudukkottai, Tamilnadu, India
Abstract
Wireless sensor networks are appealing to
researchers due to their wide range of application
potential in areas such as target detection and
tracking, environmental monitoring, industrial
process monitoring, and tactical systems. Existing
approaches are insufficient for developers to
determine whether system's requirements concern
the communication latency, bandwidth utilization,
reliability, or energy consumption. Since the data
of sensors are expressed in form of expressive
queries, the performance of query services should
be increased especially for high data rate sensor
network applications. In this paper it is decided to
propose Real Time Conflict-free Query
Scheduling (RTCQS), an enhanced transmission
scheduling technique for real time queries in
wireless sensor networks. The query preemption
algorithm is used to achieve high throughput
while scheduling without preemption is used for
the queries that are able to execute concurrent.
Keywords—Wireless Sensor Networks, throughput,
Scheduling, Interference-Communication graph, Inter
release time
I. INTRODUCTION
Sensor Networks have emerged as one of
the dominant technology trends of decade (2000-
2010) that has potential usage in defense and
scientific applications [1]. A Wireless Sensor
Network' (WSN) can be described as a network of
sensors which communicate with each other
wirelessly. These sensors may be installed in an
unattended environment with limited computation
and sensing capabilities. A typical node in the WSN
consists of a sensor, embedded processor, moderate
amount of memory and transmitter/receiver circuitry.
These sensor nodes are normally battery powered and
they coordinate among themselves to perform a
common task [2]. The ability to communicate not
only allows information and control to be
communicated across the network of nodes, but
nodes to cooperate in performing more complex
tasks, like statistical sampling, data aggregation, and
system health and status monitoring [3,4].
A. Wireless Sensor Network Protocol Stack
The communication architecture of a
wireless sensor network, implemented by a protocol
stack, is shown in Fig.1 [5] built out of 5 layers and 3
cross-layer planes. Medium access protocols reside in
the Data Link Layer, which itself not only is
responsible for a fair distribution of resources, but
also for providing data stream multiplexing, frame
detection and error control. Parts of these goals are
accomplished by MAC protocols.
The main duties of the medium access
protocol are firstly to assist the construction of a
network infrastructure, and secondly to control the
medium access, so that all sensor nodes in the
network have equal access to the resources and use
them as efficiently as possible[5]. In order to be able
to implement time-slotted MAC protocols, clock
synchronization between the sensor nodes in the
wireless network is an important requirement. This
task can be rendered possible by the Sensor
Management Protocol on the Application Layer.
B. Real Time Applications
Initial applications supported by WSNs were
mostly in environment monitoring, such as
temperature monitoring for a specific area, house
alarming, and so on. The main objectives in such
applications only involved simple data processing.
Energy consumption needed to be considered for
specific applications, so little attention was taken on
data delivery and reliability related issues [6] [7] and
[8]. WSNs have been extended and their design have
been advanced to support more hard design and
complex applications, such as security, military, fire
detection and health care related applications. In these
applications, data delivery and reliability must be
taken as important parameters in addition to energy
efficiency, because data must be collected from the
sources of events and be forwarded to the sink in real
time with high reliability, otherwise the application
will be useless. Some most important design factors
for protocols in WSN that need to be considered while
designing and deploying energy efficient MAC
protocols for any applications are the following:
network topology, type of antenna and clustering
related issues.
Sensor Networks serve many diverse
applications starting from the research on Great Duck
Island (GDI) [1] for monitoring the maine to high data
rate real time structural health monitoring, the query
services requires the improvement of performance in
terms of query throughput and latency. For meeting
this communication need, Real Time Conflict free
query scheduling (RTCQS) is designed as a WSN

2148 | P a g e
protocol which includes the common properties of
WSN query services. It provides the effective
Fig.1 Protocol Stack of WSN
prioritization between traffic classes meeting various
deadlines and supports high throughput since they
generate a very high workload. It is proposed to have
the following properties such as: 1) Adapting the
query schedule for every workload change 2)
Achieving a high throughput and low latency 3)
Ability to work on resource constrain devices 4)
Predictability in query rate and power consumption
II. RELATED WORKS
A common Medium Access Control (MAC)
paradigm used in wireless network is Carrier Sense
Multiple Access (CSMA) [9]. It is simple and flexible
not requiring any complex infrastructure, clock
synchronization or global topology. Any CSMA
based medium access scheme has two important
components, the listening mechanism and the backoff
scheme. But it doesn’t work beyond one hop causing
problem called hidden terminal problem which may
leads to degradation of throughput especially in high
data rate sensor application.
Time division multiple access (TDMA) is a
potential candidate for WSNs. It can be defined as the
process of allocating time slots to the nodes or links
between each pair of neighboring nodes, to ensure
collision free channel access. It is most attractive for
high data rate sensor networks because it is energy-
efficient and may provide higher throughput than
CSMA/CA protocols under heavy load. The two types
of scheduling carried out in TDMA are node
scheduling where the scheduler assigns slots to nodes
and link scheduling where slots are assigned to links
through which pairs of nodes communicate. Though it
is efficient than CSMA, it has many disadvantages
making less suitable for use of sensor network [10].
First, finding an efficient time schedule in a scalable
fashion is not trivial. Second TDMA needs clock
synchronization which may incur high energy
overhead because it requires frequent message
exchanges. Third, it is very expensive to handle when
sensor networks undergo frequent topology changes
because of time-varying channel conditions, physical
environmental changes, and battery outage and node
failures. Fourth, during low contention TDMA gives
much lower channel utilization and higher delays than
CSMA In addition it is not suitable for real time
applications with variable workloads as it maintains
an explicit schedule for transmission. Thus the MAC
scheme for sensor network should include a variant of
TDMA [11].
Z-MAC (Zebra MAC) is a hybrid MAC
scheme that combines the ascents of TDMA and
CSMA for sensor networks while offsetting their
descents [12]. The main feature of Z-MAC is its
adaptability to the level of contention in the network.
It is robust to dynamic topology changes and time
synchronization failures commonly occurring in
sensor networks. Z-MAC uses CSMA as the baseline
MAC scheme, but uses a TDMA schedule as a hint to
enhance contention resolution. In Z-MAC, a time slot
assignment is performed at the time of deployment
therefore a higher overhead is incurred at the
beginning. As the maximum slot number is
broadcasted in Z-MAC, security issue is to be taken
into account.
Distributed Randomized TDMA Scheduling
(DRAND) is fully distributed, efficient scalable
channel scheduling algorithm [13]. It is the first
scalable implementation of RAND which is a famous
centralized channel scheduling scheme. DRAND
calculates a TDMA schedule in time linear to the
maximum node degree in form of time slot. After the
slot assignment, each node reuses its assigned slot
periodically in every predetermined period, called
frame. A node assigned to a time slot acts as an owner
of that slot and the others be the non-owners of that
slot. It gives a chance of being more than one owner
per slot. It is useful in scheduling protocols such as Z-
MAC, FDMA, CDMA etc.,
TRaffic-Adaptive Medium Access
(TRAMA) protocol provides energy-efficient conflict
free channel access in wireless sensor networks [14].
Energy efficiency is attained by using the
transmission schedules that avoid collisions of data
packets at the receivers having nodes switch to low
power radio mode when there is no data packets
intended for those nodes. It supports for unicast,
broadcast and multicast traffic and more adaptive for
sensor network and monitoring applications. But it is
not suited for delay sensitive transmission.
Several protocols aim at supporting real-time
communication in multi-hop networks by proposing
real-time transmission scheduling for robots [15].
Both protocols may assume that at least one robot has
complete knowledge of the robots’ positions and
network topology. Though these protocols are suited
for small teams of robots, they are not suitable for
queries in multi-hop WSNs.

2149 | P a g e
In Dynamic Conflict Free Query Scheduling
(DCQS) [16], a transmission scheduling technique for
WSN queries is designed to support variable
workloads and to exploit specific communication
patterns and temporal properties of queries in WSNs.
This allows DCQS to achieve high throughput. But it
does not support query prioritization or real-time
communication.
III. SYSTEM ARCHITECTURE
RTCQS adopts a variant of node/link
scheduling called query scheduling in contrast to
earlier TDMA protocols. In query scheduling, the
time slots are assigned to transmission on specific
communication. This helps to achieve the high
throughput, low power consumption, scalability and
to adopt topology changes.
Fig.2 shows the system architecture of
RTCQS. It mainly aims at executing the real time
queries in efficient manner. For this, it is designed to
use the data aggregation functions such as packet
merging, data compression or stationary function.
Fig.2 System architecture of RTCQS
It may use the packet merging [18] as a default
aggregation function to reduce the energy
consumption and transmits workload. The two
parameters influencing the performance of query
services are throughput and real time capacity. The
former is summation of execution frequencies of all
the queries while the latter is the maximum
throughput for which the query service does not drop
packets and meets the query deadlines.
A. Components.
The two main components used in the
system are query planner and query scheduler. The
query planner decides the transmissions for each
query and scheduler for scheduling the corresponding
plans.
1) Query Planner:
The query planner is responsible for
constructing the plans which is the sequential order of
transmission steps for executing a query instance.
Each step should consist of set of transmissions which
should be free from both the primary and secondary
conflict. Primary conflict is said to occur when one
node transmits and receives at the same time slot or
receives more than one transmission destined to it at
same time slot. Secondary conflict occurs when an
intended receiver of particular transmission is also
within the transmission range of another transmission
intended for other nodes. Conflict free path can be
determined by planning the transmission with help of
Interference Communication (IC) graph.
Let G (V,E) be the IC graph where V are the
set of vertices representing sensor nodes and E are the
set of edges represents the communication edges and
interference edges. The link between the nodes for
packet transmission is termed as the Communication
edges and the link which interrupts the any
communication in time is Interference edge. For
example, PQ and RS are said to be conflict free
if PS and RS are not the edges and P, Q,
R, S are distinct. The realistic method for constructing
the IC graph is Radio Interference Detection (RID)
based on Receiver Signal Strength (RSS) [17]. An
example IC graph is shown in Fig.3
The plan with the steps 0 to 6 is shown for
the example IC graph in Fig.4. In each step the
transmissions assigned are conflict free. For example,
in step 2 the nodes v and p may transmit
simultaneously as their transmission and are
conflict free i.e, they do not conflict with each other
( ). It provides the transmission to satisfy the
precedence constraints such that t and u transmits
before its parent p.
A node may communicate with its one hop
neighbour to construct a plan at local. If the plan
involves more number of hops then the plan is
formulated by a node n with higher priority and sends
PRequest packet. Upon receiving PRequest from node
n it may check for its own one hop neighbour. If no
such node exists the receiver may respond with
PFeedback packet along with its local plan. In next
stage node n disseminates its local plan to its one hop
neighbour using PSend. On receiving the PSend
packet, it acknowledge with PCommit packet. In order
to balance the uneven workload demands across the
Sensor
network
Query
Sensor
network
Sensor nodes
Client
nodes
Base Station
Query Planner
Query Scheduler
Selecting
Conflict Free
path
Constructing the
plan for query
Evaluating Inter-release time
Priority Scheduling of the Query
IC
graph

2150 | P a g e
Fig. 3 IC graph
nodes, the planner would assign multiple steps for
these nodes to transmit to achieve their workload
demand. The minimized transmission plan may
results in reduced latency.
Fig. 4 Constructed Plan for IC graph in Fig.3
2) Query Scheduler:
As the data transmission in WSN makes use
of expressive queries, query scheduling is used to
avoid the wastage of time when the nodes or links are
idle for a period of time. Instead of assigning the time
slot for node or links, slot should be assigned for the
query. The slot is a period of time allotted for
workload on demand. Scheduling should ensure that
all the steps executed are conflict free with the relative
order being preserved.
Each node is having a local scheduler to
schedule the transmission steps. The Scheduler may
contain the start time, period of queries, plan’s length
and min. inter-release time. The min. inter-release
time is defined as the minimum time step between the
subsequent instances of query to be executed.
The algorithm for scheduling without Pre-
emption is used when the two instances of a query is
executed concurrently. The scheduler uses two
different queues called wait and execute queue. The
instance of query waiting to be executed are stored in
wait queue but are not being executed and those
instance to be executed is placed in execute queue.
The min. inter-release time is taken as between any
two instances. When no instances are executed and
the step distance between the head of wait queue and
tail of execute queue is larger than an instance of
query is started.
When an instance starts, it is moved from
wait queue to execute queue. Being simple and
efficient, it is feasible on resource constraint devices.
The operation of determining the starting time of
query instance takes time of O (1). The algorithm for
query pre-emption is used to pre-empt the instance of
query that conflict with the execution of higher
priority instance. The algorithm is shown using
Algorithm1.
begin
new instance i is released
wait = wait i
begin
start a new slot s
for each i in wait
if (may-continue(i) = true) then
continue (i)
for each i execute
run (i)
end
end
continue (i):
execute = execute i;
wait = wait – i add instances to all occurConflict
preempt(S):
execute = execute − S;
wait = wait S;
remove i from all occurConflict
may-continue(i):
if (occurConflict = null) then
return true
if (i has higher priority all instances in
occurConflict)
preempt(occurConflict);
return true
return false
run(i):
determine if node should send/recv
i = i + 1
if i = query length then execute = execute – i
occurConflict = occurConflict – i
occurConflict = occurConflict i
Algorithm1. Query Pre emption Algorithm
Senders
Steps
x y r z s q p
0 u
1 t
2 v p
3 s
4 z
5 r
6 y
y
q z
s
s
x
t
p v
r
u

2151 | P a g e
When the query instance is pre-empted, it
determines whether the query instance to be pre-
empted is conflict with existing instance. It maintains
a wait and execute queue by the priority of instances.
When a new instance is released for execution, it
should be stored in wait queue. It starts/continues an
instance i in two cases: (1) If the next step s+1 may be
executed concurrently with all instances in the execute
queue without conflict, it starts/resumes it. Then it
checks if occurConflict (set having instance conflict
with instance at execution) is empty. When an
instance is started or continued, it is moved from the
wait to execute. (2) i is also started/resumed if it has
higher priority than all the instances in occurConflict.
For i to be executed without conflict, all instances in
occurConflict must be preempted. When an instance
is preempted, it is moved from the wait to the execute
queue and it is removed from all occurConflict sets.
IV. PERFORMANCE EVALUATION
The performance of DRAND is compared
with the performance of the proposed RTCQS system
in terms of query throughput and query latency.
DRAND is a state of art protocol which does not
include any interference relationship among nodes.
Hence the results of it may produce collision. RTCQS
enforces the conflict free transmission and achieves a
high throughput and reduced latency.
Fig. 5 Performance comparison using query
Throughput for DRAND and RTCQS
A) Performance Metric. The query throughput and
query latency of the proposed system is compared
with that of DRAND which are defined as follows:
1) Query Throughput: It is the number of query
instances completed per second. It is expressed in
Hertz (Hz)
2) Query Latency: It is the response time for every
query instance after sending the query request. It is
represented by seconds (s).
B) Results.
Figure 5 shows the performance comparison of
DRAND and RTCQS in terms of throughput. Clearly
shown, RTCQS achieves the maximum throughput of
5.2Hz which is about 57% higher than DRAND. From
this it is concluded that fair allocation of slots to nodes
is unsuitable for WSN.
Figure 6 shows the performance comparison of
DRAND and RTCQS in terms of latency. Even
though the query rate
Fig. 6 Performance comparison using query latency
for DRAND and RTCQS
is low, the RTCQS performs significantly better
latency on comparing DRAND. For example, when
query rate is 3.4, DRAND has the latency of 6.4 in
contrast to RTCQS with latency of 1.5 which is about
77% lower than DRAND. The long latency period for
DRAND is due to increased waiting duration a node
to transmit entire frame to its parent.
V. CONCLUSION
The scheduling technique RTCQS is
specifically designed for real time query services in
wireless sensor networks. With the query planner,
query latency is reduced by constructing conflict free
transmission plans based on the precedence
constraints. By the query scheduler throughput is
improved by over-lapping the transmissions of
multiple query instances concurrently. Scheduler
makes use of both preemption and non preemption
technique to enable the real time applications.
Scheduler without the preemption controls only the
start of an instance; In contrast, a preemption
technique may preempt an instance to allow a higher
priority instance to execute when the two cannot be
executed concurrently. Thus RTCQS has low runtime
overhead and limited memory requirements making it
suitable for resource constrained devices and produce
high throughput. In future, the proposed technique can
0
1
2
3
4
5
6
2.5 3.4 4.3 5.2 6.1 7
QueryThroughput(Hz)
Query Rate (Hz)
Query Throughput
DRAND
RTDCQS
0
2
4
6
8
10
2.5 3.4 4.3 5.2 6.1 7
Querylatency(s)
Query rate (Hz)
Query Latency
DRAND
RTDCQS

2152 | P a g e
be applied in WSN with higher topology change to
have better performance.
ACKNOWLEDGMENT
Author express thanks to their friends and
family members for encouraging her in good manner
for completing this work.
REFERENCES
[1] I. Akyildiz, W. Su, Y. Sankarasubramaniam
and E. Cayirci: A Survey on Sensor
Networks, IEEE Communication Magazine,
pp. 102-114 (August 2002).
[2] A. Mainwaring, D. Culler, J. Polastre, R.
Szewczyk, and J.Anderson, “Wireless Sensor
Networks for Habitat Monitoring,”Proc. First
ACM Int’l Workshop Wireless Sensor
Networks and Applications (WSNA), 2002.
[3] D. Estrin, L. Girod, G. Pottie, and M.
Srivastava. “Instrumenting the world with
wireless sensornetworks”. In International
Conference on Acoustics,Speech, May 2001.
[4] D. Estrin, R. Govindan, J. S. Heidemann, and
S. Kumar.” Next century challenges:
Scalable coordination in sensor networks”. In
Mobile Computing and Networking, pages
263–270, 1999.
[5] M. A. Rahman and S. Hussain, “Optimizing
a data routing protocol’s influence on
communication connectivity,” in IEEE
Consumer Communications and Networking
Conference (CCNC’07), Jan. 11-13 2007.
[6] P. Suriyachai, U. Roedig, and A. Scott, “A
survey of mac protocols for mission-critical
applications in wireless sensor networks,”
Communications Surveys Tutorials, IEEE,
vol. PP, no. 99, pp. 1 –25, 2011.
[7] E. Felemban, C.-G. Lee, and E. Ekici,
“Mmspeed: multipath multi-speed protocol
for qos guarantee of reliability and.
timeliness in wireless sensor networks,”
Mobile Computing, IEEE Transactions on,
vol. 5, no. 6, pp. 738 – 754, June 2006.
[8] M. Soyturk and D. Altilar, “Reliable real-
time data acquisition for rapidly deployable
mission-critical wireless sensor networks,” in
INFOCOM Workshops 2008, IEEE, April
2008, pp. 1 –6.
[9] A. Woo, D. Culler, “A transmission control
scheme for media access in sensor
networks”, Proceedings of ACM
MobiCom’01, Rome, Italy, July 2001, pp.
221–235.
[10] W. Ye, J. Heidemann, and D. Estrin.
“Medium access control with coordinated
adaptive sleeping for wireless sensor
networks”, IEEE/ACM Trans. Netw.,
12(3):493{506, 2004.
[11] K. Sohrabi, J. Gao, V. Ailawadhi, G.J. Pottie,
“Protocols for self-organization of a wireless
sensor network”, IEEE Personal
Communications, October 2000, pp. 16
[12] I. Rhee, A. Warrier, M. Aia, and J. Min, “Z-
MAC: A Hybrid MAC for Wireless Sensor
Networks,” Proc. Third Int’l Conf.
Embedded Networked Sensor Systems
(SenSys), 2005.–27.
[13] I. Rhee, A. Warrior, J. Min, and L. Xu,
“DRAND: Distributed Randomized TDMA
Scheduling for Wireless Ad Hoc Networks,”
Proc. ACM MobiHoc, 2006.
[14] V. Rajendran, K. Obraczka, and J.J. Garcia-
Luna-Aceves, “Energy- Efficient Collision-
Free Medium Access Control for Wireless
Sensor Networks,” Proc. First Int’l Conf.
Embedded Networked Sensor Systems
(SenSys), 2003.
[15] A. Koubaa, M. Alves, and E. Tovar, “i-
GAME: an implicit GTS allocation
mechanism in IEEE 802.15.4 for time-
sensitive wireless sensor networks,” in
Euromicro Conference on Real-Time
Systems, 2006.
[16] O. Chipara, C. Lu, J. Stankovic, and G.
Roman, “Dynamic conflict-free transmission
scheduling for sensor network queries,”
IEEE Transactionson Mobile Computing,
vol. 10, no. 5, 2011.
[17] G. Zhou, T. He, J.A. Stankovic, and T.F.
Abdelzaher, “RID: Radio Interference
Detection in Wireless Sensor Networks,”
Proc. IEEE
[18] J. Polastre, J. Hui, P. Levis, J. Zhao, D.
Culler, S. Shenker, and I. Stoica, “A unifying
link abstraction for wireless sensor
networks,” in Embedded Networked Sensor
Systems, 2005.

Ly3421472152

More Related Content

What's hot

Viewers also liked

Similar to Ly3421472152

Recently uploaded

Ly3421472152