Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Save to My Library
Look up keyword
Like this
1Activity
0 of .
Results for:
No results containing your search query
P. 1
PHCC: Predictive Hop-by-Hop Congestion Control Protocol for Wireless Sensor Networks

PHCC: Predictive Hop-by-Hop Congestion Control Protocol for Wireless Sensor Networks

Ratings: (0)|Views: 96 |Likes:
Published by ijcsis
In wireless sensor networks (WSNs) Congestion may cause packet loss, delay, and energy waste due to a large number of packet drops and retransmissions. Therefore congestion in WSNs needs to be controlled for high energy-efficiency, to prolong system lifetime, improve fairness, and improve quality of service in terms of throughput and packet loss ratio along with the packet delay. To achieve this objective, a predictive ho-by-hop congestion control (PHCC) algorithm is proposed in this paper. The PHCC can predict congestion in the node and adjusts every upstream traffic rate with its node priority to mitigate congestion hop by hop. PHCC introduces a
priority-based rate adjustment algorithm which guarantees weighted fairness in multipath routing WSNs. In addition, PHCC can broadcast traffic on the entire network fairly. Simulation results show that the performance of proposed protocol is more efficient than previous algorithms.
In wireless sensor networks (WSNs) Congestion may cause packet loss, delay, and energy waste due to a large number of packet drops and retransmissions. Therefore congestion in WSNs needs to be controlled for high energy-efficiency, to prolong system lifetime, improve fairness, and improve quality of service in terms of throughput and packet loss ratio along with the packet delay. To achieve this objective, a predictive ho-by-hop congestion control (PHCC) algorithm is proposed in this paper. The PHCC can predict congestion in the node and adjusts every upstream traffic rate with its node priority to mitigate congestion hop by hop. PHCC introduces a
priority-based rate adjustment algorithm which guarantees weighted fairness in multipath routing WSNs. In addition, PHCC can broadcast traffic on the entire network fairly. Simulation results show that the performance of proposed protocol is more efficient than previous algorithms.

More info:

Published by: ijcsis on Jul 07, 2011
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

11/03/2011

pdf

text

original

 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 6, 2011
PHCC: Predictive Hop-by-Hop Congestion Control Protocolfor Wireless Sensor Networks
 
Shahram Babaie, Eslam Mohammadi, Saeed Rasouli Heikalabad, Hossein Rasouli
Technical and Engineering Dept., Tabriz Branch, Islamic Azad University, Tabriz, Iran
 Abstract 
 — 
In wireless sensor networks (WSNs) Congestionmay cause packet loss, delay, and energy waste due to a largenumber of packet drops and retransmissions. Thereforecongestion in WSNs needs to be controlled for high energy-efficiency, to prolong system lifetime, improve fairness, andimprove quality of service in terms of throughput and packetloss ratio along with the packet delay. To achieve this objective,a predictive ho-by-hop congestion control (PHCC) algorithm isproposed in this paper. The PHCC can predict congestion inthe node and adjusts every upstream traffic rate with its nodepriority to mitigate congestion hop by hop. PHCC introduces apriority-based rate adjustment algorithm which guaranteesweighted fairness in multipath routing WSNs. In addition,PHCC can broadcast traffic on the entire network fairly.Simulation results show that the performance of proposedprotocol is more efficient than previous algorithms.
 Keywords-wireless sensor network; congestion control; predictive; priority-based fairness
I.
 
I
 NTRODUCTION
 A typical WSN consists of a number of sensor devicesthat collaborate with each other to accomplish a commontask (e.g. environment monitoring, target tracking, etc) andreport the collected data through wireless interface to a basestation or sink node. The areas of applications of WSNs varyfrom civil, healthcare and environmental to military.Examples of applications include target tracking in battlefields [1], habitat monitoring [2], civil structuremonitoring [3], forest fire detection [4], and factorymaintenance [5].However, with the specific consideration of the unique properties of sensor networks such limited power, stringent bandwidth, dynamic topology (due to nodes failures or even physical mobility), high network density and large scaledeployments have caused many challenges in the design andmanagement of sensor networks. These challenges havedemanded energy awareness and robust protocol designs atall layers of the networking protocol stack [6].The upstream traffic from sensor nodes to the sink ismany-to-one multi-hop convergent. The upstream traffic can be classified into four delivery models: event-based,continuous, query-based, and hybrid delivery. Due to theconvergent nature of upstream traffic, congestion more probably appears in the upstream direction. Congestion thatcan leads to packet losses and increased transmission latencyhas direct impact on energy-efficiency and application QoS,and therefore must be efficiently controlled. Congestioncontrol generally follows three steps: congestion detection,congestion notification, and rate-adjusting.In response to congestion, a rate adjustment mechanismmust be designed and implemented properly in order toeliminate or avoid congestion. A number of differentschemes were reported in the literature in last few years. Astop-and-start and hop-by-hop strategy is employed in [7]. In[9] and [12], an end-to-end and AIMD-like (AdditiveIncrease Multiplicative Decrease) rate adjustment approachis employed. All of these mechanisms, however, aim atguaranteeing the simple fairness instead of the weightedfairness. A priority-based congestion control protocol(PCCP) is presented to achieve the weighted-fairnesstransmission for single-path routing WSNs in [15]. In this paper we introduce a priority-based rate adjustmentalgorithm which guarantees weighted fairness in multipathrouting WSNs. For example, this protocol can work withmultipath routing algorithm such as QEMPAR [18],concurrently.Two types of congestion could occur in WSNs [10]. Thefirst type is node-level congestion that is common inconventional networks. It is caused by buffer overflow in thenode and can result in packet loss, and increased queuingdelay. Packet loss in turn can lead to retransmission andtherefore consumes additional energy. For WSNs wherewireless channels are shared by several nodes using CSMAlike (Carrier Sense Multiple Access) protocols, collisionscould occur when multiple active sensor nodes try to seizethe channel at the same time. This can be referred to as link level congestion. Link-level congestion increases packetservice time, and decreases both link utilization and overallthroughput, and wastes energy at the sensor nodes. Bothnode-level and link-level congestions have direct impact onenergy-efficiency and QoS.Congestion control protocol efficiency depends on howmuch it can achieve the following performance objectives:(i) First, energy-efficiency requires to be improved in order to extend system lifetime. Therefore congestion control protocols need to avoid or reduce packet loss due to buffer overflow, and remain lower control overhead that willconsume additional energy more or less. (ii) Second, fairnessneeds to be observed so that each node can achieve fair throughput. Fairness can be achieved through rate-adjustment and packet scheduling (otherwise referred to asqueue management) at each sensor node. (iii) Furthermore,support of traditional quality of service (QoS) metrics such
270http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 6, 2011
as packet loss ratio and packet delay along with throughputmay also be necessary.The rest of the paper organized as follows: in section 2,we explain the related works. Section 3 describes the systemmodels. Section 4 explores the PHCC protocol with details.Section 5 describes simulation parameters and resultanalysis. Final section is containing of conclusion and futureworks.
 
II.
 
ELATED WORKS
 There are several congestion control protocols [10]-[13]for sensor networks. They differ in the way that they detectcongestion, broadcast congestion related information, and theway that they adjust traffic rate when congestion occurs. Inthis section, we review some of them and discuss their limitations.Congestion detection and avoidance (CODA) [12] proposes an open-loop, hop-by-hop backpressure mechanismand a closed-loop, multi-source regulation mechanism inevent-drivenWSNs. Sensor nodes detect congestion by monitoring thechannel utilization and buffer-occupancy level. In responseto congestion, the congested sensor nodes send backpressuremessages to their neighbors which may drop packets, reducetheir sending rate and further propagate backpressuremessages. If the sending rate of a source node is greater thanthe preset threshold, the source node must receive acontinuous stream of ACKs from the base station in order tomaintain that rate. By this means, the base station may limitthe sending rate of a source node based on deciding howmany ACKs to broadcast. CODA employs the AIMD(Additive Increase Multiplicative Decrease) coarse rateadjustment. It only guarantees simple fairness of thecongestion control.Event-to-sink reliable transport protocol (ESRT) [14]monitors the local buffer level in intermediate sensor nodesand sets a congestion notification bit in the packet when the buffer overflows. If a base station receives a packet whosecongestion notification bit is set, it broadcasts a controlsignal to inform all source nodes to reduce the sending rateaccording to certain proportion. ESRT limits sending rate of all source nodes when congestion occurs regardless of wherethe hot spot happens in WSNs. The best way is to regulatethose source nodes that are responsible for this congestion.Priority based congestion control protocol (PCCP) [15]defines a new variable, congestion degree as ratio of average packet service time over average packet inter-arrival time ateach sensor node. Congestion degree is intended to reflectthe current congestion level of each sensor node. Based oncongestion degree, PCCP employs a hop-by-hop rateadjustment technique called priority-based rate adjustment(PRA) to adjust the scheduling rate and the source rate of each sensor node in a single-path routing WSN. In the tree- based network topology of single-path routing WSNs, asensor node will only has one downstream neighbor, but itmay have multiple upstream neighbors. The whole data flowgenerated by a source node will pass through the nodes andlinks along with the single routing path. Sensor nodes learnthe number of upstream data sources in the sub tree roots andmeasure the maximum downstream forwarding rate. Finally,they calculate the per-source rate based on priority index of each source node.In Fusion [16], congestion is detected in each sensor node based on measurement of queue length. The node thatdetects congestion sets a CN (congestion notification) bit inthe header of each outgoing packet. Once the CN bit is set,neighboring nodes can overhear it and stop forwarding packets to the congested node so that it can drain the backlogged packets. This non-smooth rate adjustment couldimpair link utilization as well as fairness, although Fusionhas a mechanism to limit the source traffic rate and a prioritized MAC algorithm to improve fairness.Adaptive Rate Control (ARC), [13], is an LIMD-like(linear increase and multiplicative decrease) algorithm. InARC, if an intermediate node overhears that the packets itsent previously are successfully forwarded again by its parent node, it will increase its rate by a constant
α
.Otherwise it will multiply its rate by a factor 
β
where 0 <
β
<1. ARC does not use explicit congestion detection or explicitcongestion notification and therefore avoids use of controlmessages. However the coarse rate adjustment could result intardy control and introduce packet loss.CCF (Congestion Control and Fairness) [10] uses packetservice time to deduce the available service rate andtherefore detects congestion in each intermediate sensor node. Congestion information, that is packet service time inCCF, is implicitly reported. CCF controls congestion in ahop-by-hop manner and each node uses exact rateadjustment based on its available service rate and child nodenumber. CCF guarantees simple fairness. That means eachnode receives the same throughput. However the rateadjustment in CCF relies only on packet service time whichcould lead to low utilization when some sensor nodes do nothave enough traffic or there is a significant packet error rate(PER).Those existing congestion control protocols for WSNshave limitations. For example, they only guarantee simplefairness, which means that the sink receives the samethroughput from all nodes. However, sensor nodes may havedifferent priority or importance due to either their functionsor the location at which they are deployed. Also they work on single-path routing WSNs.III.
 
S
YSTEM
M
ODELS
 This section describes network and node models, asshown in Figs. 1 and 2, respectively.
 A.
 
 Network Model 
This paper addresses upstream congestion control for aWSN that supports multi-path routing. The network model to be investigated in this work is depicted in Fig. 1, wheresensor nodes are supposed to generate continuous data andform many-to-one convergent traffic in the upstreamdirection. Data packets are sent from nodes to sink through amulti-path and multi hop network. CSMA/CA MAC protocol is implemented in MAC layer. Each sensor nodecould have two types of traffic: source and transit. Thesource traffic is locally generated at each sensor node, while
271http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 9, No. 6, 2011
the transit traffic is from other nodes. As shown in Fig. 1,node 1 is a source node and only has source traffic, whilenodes 2, 3, 4, 5, 6 and 7 are source nodes as well asintermediate nodes because they have source traffic as wellas transit traffic. Each node could have two types of neighbor nodes: backward and forward. For example, the backwardnode of node 3 is node 1, because its data can be sent bynode 3 and forward nodes of node 3 are nodes 5, 6 and 7. Inthis paper,
 f(i)
is set of forward nodes of 
i
and
b(i)
is set of  backward nodes of 
i
. For example, in Fig.3,
b(3)
is equal {1}and
 f(3)
is equal {5, 6, 7}.
Figure 1. General network model
 B.
 
 Node Model 
 Node model of the investigated wireless sensor network is presented in Fig. 2. The source traffic of node
i
isgenerated with source traffic rate (
) by itself locally. Thetransit traffic of node
i
is received with transit traffic rate(

) from its child nodes through MAC layer of node
i
. Both
and

are converged through network layer to MAClayer as total input rate of node
i
(

). Traffic packets could be queued if 

(


) exceeds packet forwardingrate (
) at MAC layer. Congestion could take place in node
i
if 

is larger than
continuously, when the buffer of node
i
could be filled up quickly and finally overflow. Thiscongestion can be controlled by reducing

in the PHCC protocol.
Figure 2. General node model
IV.
 
PHCC
 
P
ROTOCOL
 The PHCC protocol tries to increase throughput andreduce packet loss while guaranteeing distributed priority- based fairness with lower control overhead in multi-pathrouting WSNs. The congestion control scheme of sensor node
i
is shown in Fig. 3. PHCC protocol consists of threecomponents: backward nodes selection (BNS), predictivecongestion detection (PCD) and priority-based rateadjustment (PRA), which are introduced with responsibilityfor precise congestion discovery and weighted fair congestion control.
Figure 3. Congestion control scheme in node
i
 
 A.
 
 Backward nodes selection
The node
i
selects forward nodes for itself according tomechanism used in multipath routing algorithm as well asreceived rate adjustment values from
 f(i)
. Then node
i
sendnotification to selected forward node. For increasing thethroughput, the other forward nodes of node
i
which is notselected as a forward node of this node adjust the new ratesfor their other backward nodes.
 B.
 
 Predictive congestion detection
Congestion index (
CI 
i
) reflecting the current congestionlevel at each sensor node
i
is determined on its unoccupied buffer size (
UBS 
i
) and traffic rate (
TR
i
) at MAC layer asfollows:

 

(1)

 
 
(2)

 



(3)In here,
MBS 
i
and
OBS 
i
are defined as the maximal buffer size and current queue length of node
i
.
 ji
and
ik 
denote theaverage upstream input traffic rate from node
 j
to
i
anddownstream output traffic rate from node
i
to
, respectively.On the other hand,
 ji
and
ik 
are updated periodically ateach time interval T as follows:
272http://sites.google.com/site/ijcsis/ISSN 1947-5500

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->