Multilevel priority packet scheduling scheme for wireless networks

International Journal of Distributed and Parallel Systems (IJDPS) Vol.5, No.1/2/3, May 2014
DOI : 10.5121/ijdps.2014.5307 69
MULTILEVEL PRIORITY PACKET SCHEDULING
SCHEME FOR WIRELESS NETWORKS
R.Arasa Kumar and K.Madhu Varshini
Department of Electronics and Communication Engineering, Velammal College of
Engineering and Technology, Madurai, Tamilnadu, India
ABSTRACT
Scheduling different types of packets such as real-time and non-real time data packets in wireless links is
necessary to reduce energy consumption of the wireless device. Most of the existing packet scheduling
mechanism uses opportunistic transmission scheduling, in which communication is postponed upto an
acceptable time deadline until the best expected channel conditions to transmit are found. This algorithm
incurs a large processing overhead and more energy consumption. In this paper we propose a Dynamic
Multilevel Queue Scheduling algorithm. In the proposed scheme, the ready queue is partitioned into three
levels of priority queues. Real-time packets are placed into the highest priority queue and non-real time
data packets are placed into two other queues. We evaluate the performance of the proposed Dynamic
Multilevel Queue Scheduling scheme through simulations for real-time and non-real time data. Simulation
results illustrate that the Multilevel Priority packet scheduling scheme overcomes the conventional methods
interms of average data waiting time and end-to-end delay.
KEYWORDS
Dynamic Multilevel Priority (DMP), end-to-end delay, average task waiting time.
1. INTRODUCTION
Wireless Networks have evolved a lot and there is a necessity for energy consumption of the
wireless devices and we need to manage the network resources. Wireless devices cannot have an
uninterrupted power supply. In recent wireless communication networks, the major issue is the
decrease of the transmission power consumption. The channel circumstances are time-variant in
wireless atmosphere. The Wireless channel experiences Small-scale fading due to Multipath and
large-scale fading due to Shadowing. Scheduling of packets at the data transmission part is very
much essential as it ensures deliverance of diverse data packets based on their priority. Real-time
data packets have very high priority when compared to non-real time data packets. At present,
most of the Wireless Networks operate using First Come First Served (FCFS) scheduling
algorithm that transfers the packets according to their arrival time and it needs more time to be
transmitted to a Base Station (BS). Anyway, the data packet must reach the Base Station before
the deadline or within a particular time period. And also within minimum end-to-end delay, the
real-time data must be delivered to BS.
In this paper, we propose Dynamic Multilevel Queue Priority Scheduling algorithm where the
nodes are organized into a hierarchical structure. Nodes which have equal hop distance from the
BS are said to be present at the same hierarchical level. Each node upholds three levels of Priority
Queues, since we organize data packets as a) real-time (Priority-1) b) non-real time data packets

70
from lower level nodes (Priority-2) c) non-real time data packets present at the node itself
(Priority-3). Shortest Job First (SJF) scheduler is being used to process the non-real time data
packets that are present at the same level priority.
The remaining of this paper is organized as follows. Section II discusses various Related Works.
Section III presents a review of numerous existing Packet Scheduling algorithms. Section IV
describes various postulations and terminologies of DMP packet scheduling scheme. Section V
presents the experimental results and finally Section VI concludes the paper.
2. RELATED WORKS
The energy consumption problem in Wireless Networks has attracted world-wide. Many works
[1-3] has been done with a vision of minimizing the energy consumption on the wake-up mode in
wireless systems. A real-time architecture for large-scale networks [26] was proposed where
priority based scheduler is used. The data packets which travel maximum distance from the
source node to BS and have the minimum deadline are prioritized. A packet scheduling scheme
and algorithm called RACE [27] for real-time large scale networks was proposed. It uses
Bellman-Ford algorithm inorder to find out ways with less traffic and delay. Earliest Deadline
First (EDF) scheduling algorithm was used in RACE to transmit packets with shortest deadline.
[29] presents the mostly used operating system of Wireless Network and differentiate them as Co-
operative and Preemptive. Co-operative scheduling algorithms are based on Adaptive Double
Ring Scheduling (ARDS) and EDF [30], that has two queues with various priorities. Based upon
the deadline of the arriving packets , the scheduler switches between the two queues. Co-
operative schedulers are used in applications with limited resources. Preemptive Scheduling is
based on EFRM scheme which is an extension of Rate Monotonic (RM) scheme. In [6] , the state
of distributed data aggregation in Wireless Networks is being reviewed.
3. REVIEW OF VARIOUS SCHEDULERS
In this Section, various conventional packet scheduling algorithms are discussed.
3.1 FACTOR: DEADLINE
Packet Scheduling algorithms are classified depending on the deadline of the arrival of data
packets to the Base Station (BS).
3.1.1. First Come First Served (FCFS): Most of the existing Wireless applications use First
Come First Served Schedulers in which the datas are processed according to their arrival times at
the ready queue. Here, data from the distant nodes which comes later at the intermediate nodes
need more time to be delivered to the Base Station (BS) but packets from the nearby nodes take
less time at the intermediate nodes. In FCFS, most of the packets practice longer waiting time.
3.1.2. Earliest Deadline First (EDF): Whenever there are more data packets present at the ready
queue and those packets have a deadline within which it must be transmitted to BS, the packet
which has the earliest deadline is transmitted first. This is said to be as an efficient algorithm
interms of end-to-end delay and average packet waiting time.

71
3.2 FACTOR: PRIORITY
Packet Scheduling algorithms are classified according to the priority of data packets.
3.2.1.Non-Preemptive: In Non-preemptive packet scheduling algorithm, when a packet P1 starts
processing, process p1 carries on even if a higher priority data packet p2 arrives at the ready
queue. Thus p2 must wait in the ready queue till the completion of the process p1.
3.2.2Pre-emptive: In Preemptive packet scheduling algorithm, the context of lower priority
packets is saved by processing the higher priority packets first.
3.3 FACTOR: PACKET TYPE
On the basis of data packet types , packet scheduling algorithms are divided as
3.3.1 Real-time packet scheduling : Based on the priority and packet types, packets at the nodes
must be scheduled. Amid all the data packets in the queue, real-time data packets are regarded as
the highest priority packets. Thus the real-time emergency are processed first and then transmitted
to the Base Station with minimum end-to-end delay.
3.3.2 Non-real time packet scheduling: Non-real time data packets have lesser priority when
compared to real-time data packets. In scheduling of non-real time data packets either First Come
First Served (FCFS) or Shortest Job First (SJF) scheduling algorithm can be used at the ready
queue of each node.
3.4. FACTOR: NUMBER OF QUEUES
On the basis of number of levels of a node, packet scheduling algorithms are classified as
3.4.1. Single-Queue: Every node has a ready queue. Data packets of all the types reach the ready
queue and are scheduled on the basis of size, type, priority, etc., This type of scheduling has high
starvation rate.
3.4.2. Multi-level Queue: A node has two or more queues. Packets are kept inside the queues
based on their types and priorities. The ready queue gets separated into three levels of priorities.
Real-time data packets with highest priority is kept in first priority queue and is processed using
FCFS. Non-real time data packets are put into the lower second and third priority levels and
processed using different scheduling algorithms. Data packets are scheduled in each queue or
among different queues. A node at the lowest level has lesser number of queues whereas a node at
the higher level has many queues inorder to minimize the end-to-end transmission delay and
maintain energy consumption in the network.
4. DYNAMIC MULTILEVEL PRIORITY PACKET SCHEDULER
4.1 POSTULATIONS
We consider the following postulations to implement DMP scheduling algorithm in this section.
Data packets consists of either real-time data or non-real time data. All the data packets arriving

72
at the queue are of equal size. Data aggregation is not carried out at the intermediate nodes for
real-time packets. Nodes are assumed to be situated at different levels on the basis of number of
hops from Base Station.
Time Division Multiple Access scheme is being employed to allot timeslots to nodes at various
levels. Nodes at the lowest level are allotted timeslot1. There are three levels of priority in the
ready queue such as real-time data (priority 1) and Non-real time data (priority 2 and 3).
The length of the priority level queues is variable. Priority 1 level queue is shortest. The length of
both priority 2 and priority 3 level queues is same.
4.2 TERMINOLOGIES
4.2.1.Routing Protocol: We use Zone-based routing protocol for better energy efficiency and for
stability in energy consumption among the nodes. In a zone-based routing protocol, each zone
contains a Zone Head (ZH) for identification and their structure is based upon the number of hops
they are far-off from the Base Station (BS). Nodes in the zones that are one hop distant from the
BS is considered to be at level 1 and similarly those which are at two hop distant from the BS is
considered to be at level 2. Each zone is subdivided into smaller number of squares such that a
node in square envelops all other neighboring squares.
4.2.2.TDMA Systems: Task scheduling at each level is performed using TDMA method. Data
packets are sent from the lower level nodes to the BS via intermediate nodes. Comparing to the
lower-level nodes , nodes present at the top and intermediate levels have more processing
conditions. The time-slot of lower-level nodes is set to lowest length contrast to the higher value
of length of the timeslot of upper-level nodes. Intermediate level nodes must be stopped from
aggregating data as they have to be transmitted to the users with possibly lesser delay.
4.2.3. Fairness: This metric guarantees that tasks of diverse priorities are done with least waiting
time on the basis of priority of chore at the ready queue. For example, if any lower-priority chore
waits for more time for the constant income of higher-priority chores, fairness defines a limit that
permits the lower-priority tasks to be processed following a definite waiting time.
4.3. WORKING PRINCIPLE
Data packets that arrive at a node are scheduled amongst all the levels in the ready queue. Next,
data packets in each level of the queue are scheduled. Each node at various levels consists of a
variable length ready queue. Pr1 queue is meant for real-time data, Pr2 queue is for non-real time
remote data and Pr3 queue is for non-real time local data. The data packets from the lowest level
nodes traverses various intermediate nodes and finally reaches the BS. The proposed scheduling
method presumes that the nodes are nearly organized in a hierarchical structure. Nodes which are
at the equal hop count from the Base Station (BS) are regarded as situated at the same level.
Time-Division Multiplexing Access is being used for the processing of data packets at different
levels. For example, nodes that are situated at the lowest level and the immediate next lowest
level can be allotted timeslots 1 and 2 respectively.
We take the largest number of levels in the queue of a node to be three. The motive for selecting
maximum three number of queues are

73
i. Real-time emergency packets with highest priority to accomplish the overall aim of the
Wireless Networks
ii. Non-real time packets to accomplish minimum average waiting time and end-to-end delay
iii. Non-real time packets with lowest priority to accomplish fairness.
5. PERFORMANCE ESTIMATION
The simulation model has been executed by using MATLAB. In DMP packet scheduling scheme,
one of the priority levels for non-real time data packets use Dynamic RR scheduling algorithm.
This is used to reduce average waiting time and end-to-end delay. The performance is evaluated
for the DMP packet scheduling algorithm, contrast to Dynamic Round Robin scheduling concept.
This comparison is done in terms of end-to-end transmission delay and average packet waiting
time. The number of simulated zones differs from 4 to 12 zones. Nodes are allocated identically
over the zones. The ready queue at each node can hold utmost 50 tasks. Type ID is used inorder
to identify its type.
Figure1. End-to-End delay over a number of zones

74
Figure 2. End-to-End delay over a number of levels
Figure 3. Waiting time of real-time data over a number of zones
Fig 1 & 2 demonstrate the end-to-end data transmission delay of real-time tasks over a number of
zones and levels, respectively. In both cases, we examine that DMP using Dynamic RR
outperforms the traditional DMP packet scheduler. This is due to the highest priority given to the

75
real-time tasks and since it allows the real-time packets to preempt the process of non-real time
data packets. Real-time packets have lesser transmission delay.
Fig 3 illustrates that the DRR packet scheduler has better performance compared with the DMP
packet scheduler interms of average task waiting time.
6. CONCLUSION
In this paper, we discuss about Dynamic Multilevel priority packet scheduling method. This type
of scheduler uses three priority levels. Here, the ready queue gets separated into three levels of
priority queues. Higher priority has been given to the processing of real-time data packets. In our
proposed method, we modify one of the scheduling scheme for processing non-real time data
packets and we apply Dynamic RR scheme in the second or third priority level of the Dynamic
Multilevel queue. This is done in order to minimize the average waiting time of the data packets
and also minimizes end-to-end delay. Experimental results show that the DMP using DRR
outperforms the normal DMP scheme interms of end-to-end delay and average data waiting time.
REFERENCES
[1] R. Zheng, J. C. Hou, and L. Sha, “Asynchronous wakeup for ad hoc networks,” in Proc. 4th ACM Int.
Symp. Mobihoc, Annapolis, MD,Jun. 2003, pp. 35–45.
[2] X. Yang and N. H. Vaidya, “A wakeup scheme for sensor networks: Achieving balance between
energy saving and end-to-end delay,” in Proc.IEEE 10th RTAS, Toronto, Canada, May 2004, pp. 19–
26.
[3] M. Peng and W. Wang, “An adaptive energy saving mechanism in the wireless packet access
network,” in Proc. IEEE WCNC, Las Vegas, NV, Apr. 2008, pp. 1536–1540.
[4] R. Yu, Y. Zhang, Z. Sun, and S. Mei, “Energy- and QoS-aware packet forwarding in wireless sensor
networks,” in Proc. IEEE ICC, Glasgow, U.K., Jun. 2007, pp. 3277–3282.
[5] A. O. Fapojuwo, “Energy consumption and message delay analysis of QoS-enhanced base-station-
controlled dynamic clustering protocol for wireless sensor networks,” IEEE Trans. Wireless
Commun., vol. 8, no. 10, pp. 5366–5374, Oct. 2009.246810121400.20.40.60.811.21.41.61.82x 104No
of zonesReal time Task End to End delay DMPDRR
[6] Z. Ye, A. A. Abouzeid, and J. Ai, “Optimal stochastic policies for distributed data aggregation in
wireless sensor networks,” IEEE/ACM Trans. Netw., vol. 17, no. 5, pp. 1494–1507, Oct. 2009.
[7] G. Anastasi, M. Conti, and M. Di Francesco, “Extending the lifetime of wireless sensor networks
through adaptive sleep,” IEEE Trans. Industrial Informatics, vol. 5, no. 3, pp. 351–365, 2009.
[8] G. Bergmann, M. Molnar, L. Gonczy, and B. Cousin, “Optimal period length for the CQS sensor
network scheduling algorithm,” in Proc. 2010 International Conf. Netw. Services, pp. 192–199.
[9] P. Guo, T. Jiang, Q. Zhang, and K. Zhang, “Sleep scheduling for critical event monitoring in wireless
sensor networks,” IEEE Trans. Parallel Distrib. Syst., vol. 23, no. 2, pp. 345–352, Feb. 2012.
[10] F. Liu, C. Tsui, and Y. J. Zhang, “Joint routing and sleep scheduling for lifetime maximization of
wireless sensor networks,” IEEE Trans.Wireless Commun., vol. 9, no. 7, pp. 2258–2267, July 2010.
[11] J. Liu, N. Gu, and S. He, “An energy-aware coverage based node scheduling scheme for wireless
sensor networks,” in Proc. 2008 International Conf. Young Comput. Scientists, pp. 462–468.
[12] O. Khader, A. Willig, and A. Wolisz, “Distributed wakeup scheduling scheme for supporting periodic
traffic in wsns,” in Proc. 2009 European Wireless Conf., pp. 287–292.
[13] B. Nazir and H. Hasbullah, “Dynamic sleep scheduling for minimizing delay in wireless sensor
network,” in Proc. 2011 Saudi International Electron., Communications Photon. Conf., pp. 1–5.
[14] D. Shuman and M. Liu, “Optimal sleep scheduling for a wireless sensor network node,” in Proc. 2006
Asilomar Conf. Signals, Syst. Comput., pp. 1337–1341.

76
[15] S. Paul, S. Nandi, and I. Singh, “A dynamic balanced-energy sleep scheduling scheme in
heterogeneous wireless sensor network,” in Proc. 2008 IEEE International Conf. Netw., pp. 1–6,
2008.
[16] A. R. Swain, R. C. Hansdah, and V. K. Chouhan, “An energy aware routing protocol with sleep
scheduling for wireless sensor networks,” in Proc. 2010 IEEE International Conf. Adv. Inf. Netw.
Appl., pp. 933–940.
[17] N. Edalat, W. Xiao, C. Tham, E. Keikha, and L. Ong, “A price-based adaptive task allocation for
wireless sensor network,” in Proc. 2009 IEEE International Conf. Mobile Adhoc Sensor Syst., pp.
888–893.
[18] H. Momeni, M. Sharifi, and S. Sedighian, “A new approach to task allocation in wireless sensor actor
networks,” in Proc. 2009 International Conf. Computational Intelligence, Commun. Syst. Netw., pp.
73–78.
[19] W. Stallings, Operating Systems, 2nd edition. Prentice Hall, 1995.
[20] F. Tirkawi and S. Fischer, “Adaptive tasks balancing in wireless sensor networks,” in Proc. 2008
International Conf. Inf. Commun. Technol.: From Theory Appl., pp. 1–6.
[21] X. Yu, X. Xiaosong, and W. Wenyong, “Priority-based low-power task scheduling for wireless sensor
network,” in Proc. 2009 International Symp. Autonomous Decentralized Syst., pp. 1–5.
[22] Y. Zhao, Q. Wang, W. Wang, D. Jiang, and Y. Liu, “Research on the priority-based soft real-time
task scheduling in TinyOS,” in Proc. 2009 International Conf. Inf. Technol. Comput. Sci., vol. 1, pp.
562–565.
[23] TinyOS. Available: http://webs.cs.berkeley.edu/tos, accessed June 2010.
[24] Available: http://webs.cs.berkeley.edu/tos, accessed June 2010.
[25] E. M. Lee, A. Kashif, D. H. Lee, I. T. Kim, and M. S. Park, “Location based multi-queue scheduler in
wireless sensor network,” in Proc. 2010 International Conf. Advanced Commun. Technol., vol. 1, pp.
551–555.
[26] 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,” in Proc. 2002 IEEE Real-Time Embedded
Technol. Appl. Symp., pp. 55–6
[27] K. Mizanian, R. Hajisheykhi, M. Baharloo, and A. H. Jahangir, “RACE: a real-time scheduling policy
and communication architecture for largescale wireless sensor networks,” in Proc. 2009 Commun.
Netw. Services Research Conf., pp. 458–460.
[28] M. Yu, S. J. Xiahou, and X. Y. Li, “A survey of studying on task scheduling mechanism for TinyOS,”
in Proc. 2008 International Conf. Wireless Commun., Netw. Mobile Comput., pp. 1–4.
[29] P. A. Levis, “TinyOS: an open operating system for wireless sensor networks (invited seminar),” in
Proc. 2006 International Conf. Mobile Data Manag., p. 63.
[30] K. Lin, H. Zhao, Z. Y. Yin, and Y. G. Bi, “An adaptive double ring scheduling strategy based on
tinyos,” J. Northeastern University Natural Sci., vol. 28, no. 7, pp. 985–988, 2007.
Authors
R.Arasa Kumar was born on 15.11.1985. He completed his B.E. in Electronics and
Communication Engineering in 2007 and his M.E. in Communication Systems in
2009. He has teaching experience of about 5yrs. At present, he is working as an
Assistant Professor in Velammal College of Engineering and Technology, Madurai,
Tamilnadu, India. He has published 3 papers in International Journals. His area of
interest includes Wireless Communication.
K.Madhu Varshini completed her B.E. in Electronics and Communication in 2012 and
now doing her final year M.E in Velammal College of Engineering and Technology,
Madurai. Her areas of interest include Wireless Networks and Optical
Communication. She has published a paper in International Conference.

Multilevel priority packet scheduling scheme for wireless networks

More Related Content

What's hot (17)

Similar to Multilevel priority packet scheduling scheme for wireless networks (20)

Recently uploaded (20)

Multilevel priority packet scheduling scheme for wireless networks