Sie sind auf Seite 1von 6

Fairness for Extend DCCP Congestion Control in Wireless Sensor Networks

LIU Yong-Min 1, 2, JIANG Xin-Hua 2, NIAN Xiao-Hong 2


1. College of Vocational Technology, Central South University of Forestry and Technology, Changsha 410004, China
E-mail: lym37212004@126.com
2. School of Information Science and Engineering, Central South University, Changsha 410075, China
E-mail: lym37212002@yahoo.com.cn

Abstract: This TCP is primarily designed for wired networks and became very efficient and robust with years of
enhancements. However, experiments and research showed that TCPs congestion control algorithm performs very
poorly over Wireless Sensor Networks (WSN) with severe unfairness among flows. Because the deployment of a sensor
network causes unpredictable patterns of connectivity and varied node density, resulting in uneven bandwidth
provisioning on the forwarding paths. This paper studies TCPs fairness issues in WSNs, and designs an Extended
congestion control algorithm based on the characteristics of the WSNs. The protocol is designed as extension to DCCP
(Datagram Congestion Control Protocol) with a new congestion control component. We also implemented this
congestion control algorithm in NS2. Simulation results show improvements on fairness achieved by using Extend
congestion control algorithm.
Key Words: Max-min Fairness, Datagram Congestion Control Protocol, congestion control, wireless sensor networks

IEEE 802.11 wireless links, the work here also assumes that
1 INTRODUCTION the MAC layer is an IEEE 802.11-like random access
WSN typically consists of a large number of tiny wireless protocol.
sensor nodes (often referred to as nodes or motes) that are
densely deployed [1]. Nodes measure some ambient
conditions in the environment surrounding them. These
measurements are, then, transformed into signals that can
be processed to reveal some characteristics about the
phenomenon. The data collected is routed to special nodes,
called sink nodes (or Base Station, BS), typically in a
multi-hop basis. Then, the sink node sends data to the user.
Depending on the distance between the user and the
network, a gateway may be needed in order to bridge both,
either through the Internet or satellite. Two sensors are
neighbors if they can directly communicate with each other.
As is shown in the figure 1. Fig 1. Wireless Sensor Networks.
Sensor networks have a wide range of applications in
habitat observation, health monitoring, object tracking, The sensors share the same wireless media and each packet
battlefield sensing, etc. They are different from traditional is transmitted as a local broadcast in the neighborhood. We
wireless networks in many aspects [2,3]. Particularly, assume the existence of a MAC protocol, which ensures
sensor nodes are limited in computation capability, memory that, among the neighbors in the local broadcast range, only
space, communication bandwidth, and above all, energy the intended receiver keeps the packet and the other
supply. neighbors discard the packet. The sensors are statically
located after deployment. We study data packets sent from
Nowadays, nodes are intended to be small and cheap. sensors to base stations. The base stations are connected via
Consequently, their resources are limited (typically, limited an external network to a data collection center. A data
battery, reduced memory and processing capabilities). packet may be sent to any base station as long as there is a
Because of the restrained transmission power, wireless forwarding path.
sensor nodes can only communicate locally, with a certain
number of local neighbors. So, nodes have to collaborate in Transmission Control Protocol (TCP) is a reliable,
order to accomplish their tasks: sensing, signal processing, end-to-end transport protocol, which is widely used for data
computing, routing, localization, security, etc. services and is very efficient for wired networks. However,
Consequently, WSN are, by its nature, collaborative experiments and research showed that TCPs congestion
networks [4]. As most wireless Networks are built based on control algorithm performs very poorly over Wireless
sensor networks with degraded throughputs [5]. Research
therefore has focused on further improving TCP to address
This project supported by National Natural Science Foundation of P. R. the special characteristics of Wireless sensor networks.
China under Grant (60474029, 60774023, 60774045), China Postdoctoral
Science Foundation under Grant (2005038558).

978-1-4244-2723-9/09/$25.00 
c 2009 IEEE 4732
Currently, the vast majority of the traffic in the Internet the worst rate in the most congested area. Directed
relies upon the congestion control mechanism provided by Diffusion [9] and SPEED [10] were not specifically
TCP. However, applications such as streaming video and designed for congestion control, but they may be adapted
Internet Telephony prefer timeliness to reliability. The for this purpose to a certain degree.
reliability and in-order delivery algorithm provided by TCP
often results in arbitrary delay, and TCPs rate control 2.2 TCP Fairness
AIMD (Additive Increase and Multiplicative Decrease) TCP flows experience severe unfairness in wireless sensor
algorithm causes very sharp bandwidth change upon the networks. TCPs window-based congestion control adjusts
detection of one packet loss. Consequently, such the congestion window size every RTT. Flows with longer
applications often choose UDP, with either their own RTT increase the congestion window slower than flows
congestion control mechanisms implemented on top of it or with shorter RTT. At the network routers, an unfair
none at all. Long-lasting UDP flows without any packet-dropping scheme, such as a simple FIFO drop tail
congestion control mechanism present a potential threat to scheme, may cause some flows to experience more losses
the network. Also, congestion control mechanisms are than others. Medium access at a gateway is inherently
difficult to implement and may behave incorrectly. unfair when using a MAC protocol such as IEEE 802.11.
A sensor is congested if it receives more traffic than its Upstream flows (from senders to the gateway) tend to
maximum forwarding rate. The nature of sensor occupy the whole media and the downstream flows (from
deployment leads to unpredictable patterns of connectivity the gateway to receivers) almost stop transmission when
and varied node density, which causes uneven bandwidth multiple upstream and downstream flows co-exist.
provisioning on the forwarding paths. The data sources are Unfairness between the upstream and downstream flow
often clustered at sensitive areas under scrutiny and may throughputs is extremely high, with a ratio of up to 800
take similar paths to the base stations. When data converge between them [11]. In a Wireless sensor network, IN TCP
toward a base station, congestion may occur at sensors that flows (from the wired part to the wireless part) get more
receive more data than they can forward. bandwidth than the coexisting OUT TCP flows (from the
Congestion causes many problems. When a packet is wireless part to the wired part) [12]. IN flows obtain a much
dropped, the energy spent by upstream sensors on the higher share of the bandwidth when mixed flows exist due
packet is wasted. The further the packet has traveled, the to exposed and hidden node effects. TCPs own timeout and
greater the waste. When a sensor X is severely congested, if back-off schemes further worsen the unfairness.
the upstream neighbors attempt to send to X, their efforts The max-min flow control was first proposed by Faffe [13]
(and energy) are deemed to be wasted and, worse yet, to distribute the network bandwidth fairly among a set of
counter-productive because they compete for channel best effort flows. The name max-min comes from the
access with neighboring sensors. Finally, and above all, the strategy of maximizing the bandwidth allocated to those
data loss due to congestion may jeopardize the mission of flows that receive the minimum bandwidth. Much further
the application. While fusion techniques [6] can be used for research [14,15] has been done since then. All these works
data aggregation, applications may require some specifics assume that each flow has a fixed routing path. Two basic
(e.g., exact locations of the reporting sensors) to be kept [1], properties of the max-min flow control are:
which place a limit on how much the fusion can do. z Fairness property.
At each link, any passing flow is entitled to an equal share
2 RELATED WORK
of the link capacity unless the flow is limited to a smaller
bandwidth at another link on its path.
2.1 Congestion control
z Maximum throughput property.
TCP provides a connection-oriented, reliable data The entire capacity of a link must be allocated to the flows
transmission. The basic idea of TCP congestion control is unless every passing flow has a bottleneck link elsewhere
that TCP senders probe the network for available resources, which limits the bandwidth that the flow can receive.
and increase the transmission rate until packet losses are
A bottleneck algorithm that assigns the max-min bandwidth
detected. TCP takes packet loss as indication of network
to every flow was described in [16, 17] and is repeated here:
congestion and triggers appropriate congestion control
Find the global bottleneck link that has the smallest
schemes.
bandwidth per flow. Assign an equal share of the links
The problem of congestion control in sensor networks is capacity to each passing flow. Remove the link and the
largely open. A typical approach is for a congested sensor passing flows from the network. When a flow is removed,
to send backpressure messages to its neighbors [7], which the capacities of all links on its routing path are reduced by
reduce their data rates and may further propagate the the bandwidth assigned to the flow. Repeat the above
backpressure messages upstream. However, the important process until every flow is assigned a bandwidth and
issue of ensuring fairness among the sensors during their removed from the network.
rate reduction is not addressed by this approach.
In general, TCP works poorly in Wireless sensor Networks.
In ESRT [8], by monitoring the congestion notification bit This is caused by the high bit error rate over wireless links,
carried in the packet header, the base station decides a as well as TCPs built in congestion control algorithm
common rate for all sensors such that no packet will be lost working with the contention based media access of IEEE
in the network. This approach achieves fairness but is too 802.11. A large amount of research has focused on
pessimistic because every sensor must conform its rate to improving the fairness issues discussed above.

2009 Chinese Control and Decision Conference (CCDC 2009) 4733


2.3 TCP Fairness improvements connections with a long propagation delay, such a small
window size will affect the efficiency.
A wireless network has different properties than a wired
network. For example, the capacity of a wireless link To improve fairness over a combined wired and sensor
between two neighbors is not fixed but depends on the network, a non-work-conserving scheduling algorithm
amount of background communication in the working with IEEE 802.11 MAC is proposed in [24]. In the
neighborhood. MAC-layer fairness in wireless networks proposal, the normal FIFO work conserving scheduling
was studied in [18,19]. Flow-level proportional fairness in scheme is replaced, which treats routing packets (generated
FDMA/CDMA networks was studied in [20] under the by routing protocols) as high priority packets over data
assumption that each flow has a single routing path. packets (generated by applications), and puts the high
Max-min fairness among one-hop flows in FDMA/CDMA priority packets in the queue before all data packets upon
networks was studied in [21]. Fairness in TDMA networks arrival. The head of the queue is sent to the MAC after
was studied in [22] under the assumption of a tree routing knowing that the MAC is ready to send another packet. A
structure from all data sources to a sink. The TAP fairness timer is set after a data packet is sent to the MAC. Only
in wireless backhaul networks was investigated in [23], after the timer expires can the queue send another data
which achieves temporal fairness instead of throughput packet. The routing packets have high priority and dequeue
fairness. However, the paper does not provide an algorithm immediately after knowing that the MAC is ready. No timer
that computes the TAP rates. is set after a routing packet is sent. The duration of the timer
is based on the queue output rate and is the sum of three
Several researchers have studied TCP fairness. In [11], TCP
parts: transmission delay without contention; transmission
unfairness among upstream and downstream flows is
delay based on recent queue output (choosing from four
demonstrated and investigated. A gateway is used to
predefined values based on the queue output rate); and a
forward traffic, and the buffer size in the gateway plays a
random value uniformly distributed from zero to the value
key role in obtaining fair sharing of the medium among
of the second part. The timer adds extra adaptive delay in
upstream and downstream flows. [11] shows via simulation
the scheduling, so the more aggressively a node is sending
that, when equal number of downstream and upstream
packet, the more severely it is penalized, thereby nodes
flows exist, the average throughput ratio between the
failing to grab the medium can compete with the fast
upstream and downstream flows can go up to 800. The
sending nodes now.
reason is that upstream flows ACKs clutter the gateway
buffer and cause the buffer to overflow. Downstream flows Through simulations, [24] shows that the severe unfairness
experience timeouts and transmit only with a window of among flows can be eliminated while the aggregate
0-2 packets because of the packet drops at the gateway throughput experiences a small degradation. Also, the
buffer. Upstream flows normally can reach their maximum maximum congestion window size does not adversely
window size. Because of the cumulative nature of TCP impact fairness in this scheme, so unlike the previous
ACKs, small losses of ACKs do not affect the window size. schemes there is no need to pre-configure the maximum
congestion window size or to modify the advertised
The proposed solution is to advertise the available buffer
receiver window.
size to the sender. The gateway keeps the number of current
TCP flows in the system. If the buffer size at the gateway is It is well known that proportional or temporal fairness is
B and the number of flows is N, then the receiver window more appropriate in a multi-rate wireless network [23,25],
of all the TCP flows are set to the minimum of advertised where max-min fairness may cause severe throughput
receiver window or [B/N] by modifying the receiver degradation. That is NOT the case for a single-rate wireless
window field of ACKs traversing the gateway. Through sensor network, which is the subject of this paper. We study
simulation and test bed implementation, this proposal the sensor networks whose transceivers operate at a single
shows a very good fairness, with the throughput ratio of transmission rate.
upstream and downstream flows being 1 in the simulation In summary, a range of proposals has addressed how to
and 1.007 in the test bed. The study is based on the increase TCP fairness, with varying degrees of success.
assumption that all the losses happen in the gateway due to However, these approaches are all limited by their intent to
buffer overflow and all RTTs are the same among flows. keep at least the TCP semantics unchanged, if not the TCP
In [12], the TCP fairness problem in a combined wireless implementations at each node, often resulting in
and wired network is investigated. The study shows that IN improvement in one aspect (such as throughput) while
flows get significant more bandwidth than OUT flows. This trading off another aspect (such as fairness). Also, some of
unfairness is the joint result of the MAC layers exposed the proposals are only applicable in pure wireless nodes.
nodes and hidden nodes problem and TCPs timeout and However, we believe that the more relevant network
back-off schemes. In a study performed on the test bed, it is architecture is Wireless Sensor Networks. Finally, none of
found that when the maximum congestion window size is these proposals will address the congestion control problem
smaller than a certain value (8 in the test), the two flows for streaming UDP flows in such networks. In the next
share the bandwidth fairly and the aggregate throughput section, we discuss a new DCCP protocol that improves on
reaches the upper limit. The problem is that this window fairness (compared to TCP), and can be suitable for both
size could not be preconfigured. A similar study is reliable data transfer and streaming media flows.
conducted in a pure sensor network, and the optimal
congestion window size is found to be 1-2 packets. For

4734 2009 Chinese Control and Decision Conference (CCDC 2009)


3 Extended DCCP Congestion Control corrupted packets by the senders, detection and deletion of
duplicated packets at the receivers, and in-order delivery of
3.1 Standard DCCP received packets to the application program at the receivers.
In extended protocol, the sender has four states: Normal
If a sensor receives more data than it can forward, State, Congestion State, Failure State (route change or
congestion will occur and the sensor will have to drop the link failure) and Error State (transmission error). Rate
excess packets. In order to avoid congestion, the upstream based congestion control is used to avoid the frequent slow
sensors must redirect packets to other paths. If all starts. The most important task is to design the rate equation
forwarding paths from an active sensor to the base stations for each state, which is the key for fairness.
are congested, the sensor must generate data at a reduced
rate. The problem becomes interesting when many active To determine the available end-to-end bandwidth, we
sensors share forwarding paths in an arbitrary way. For adopted the delay based rate estimation mechanism in
each active sensor, we want to find the highest possible rate FAST TCP [27]. The sender maintains two RTT values,
that does not cause congestion at a downstream node. We one is base RTT (baseRTT), which is the minimum
also want all active sensors to have equal access to the recorded RTT, and the other is exponentially averaged RTT
transmission capacity of the network, no matter how (avgRTT). Each time the sender goes into the failure state,
different their forwarding paths are. the baseRTT will be reset by the round trip time of a probe
packet and its corresponding acknowledgment, after being
temporarily saved as old baseRTT. The sending rate after
the route establishment is proportional to baseRTT/Old
baseRTT.
In the Normal State, the sender adjusts the rate proportional
to baseRTT/avgRTT.
Fig 2. The architecture of DCCP. In the Congestion State, when ECN mark without packet
loss happened, the rate adjustment is the same as in Normal
The purpose of DCCP is to provide a standard way to State. But when packet loss happened, the sending rate will
introduce congestion control and congestion control halve. This idea is based on FAST TCP for High-Speed
negotiations into multimedia applications, the Internet Long-Distance Networks, which showed proportional
Engineering Task Force (IETF) defined the Datagram fairness under no congestion or mild congested situations
Congestion Control Protocol [26] in 2006, as an alternative when packet loss occurs infrequently.
to UDP for unicast multimedia applications that prefer
timeliness of data to reliability. This new transport protocol In the Failure State, probe packets are send out to monitor
is designed for deployment as a standard feature in end the network situation. The rate of sending probe packets can
hosts (PCs, VoIP codecs, and other internet-enabled be set to one packet per RTO like in Fixed RTO, but it
multimedia appliances). This is to say, DCCP [26] is a new should be studied further by experiments.
protocol designed for applications that require the In the Error State, the rate is set to *rate, calculated using
flow-based semantics of TCP, but prefer timely delivery to the above scheme, where  ranges from 0.5 to 1, according
in-order delivery, or a congestion control mechanism to the error rate.
different from what TCP provides. DCCP aims to be a A simplified DCCP with rate-based congestion control is
minimal overhead and general-purpose transport-layer implemented based on the TCP implementation in NS2.
protocol providing only two core functions: The Because wireless Sensor nodes do not support ECN and the
establishment, maintenance and teardown of an limitation of getting network detected link failure in NS2,
unreliable packet flow, and Congestion control of that the implementation has only two states: Congestion State
packet flow. The architecture of DCCP is shown in the and Normal State.
figure 2. In the implementation, ACKs are sent back to the sender
whenever the receiver receives a packet. ACKs have the
3.2 Extended DCCP Scheme ACK Vector option as specified in the DCCP specification.
To improve the performance of WSNs throughout, the ACK vectors contain packet reception information
Extended DCCP was proposed in this section. It utilizes (whether they are received, not received or ECN marked).
DCCP with the congestion control mechanism specified in Also, the ACK Vector can be used to return information
a new Congestion Control Identifier (CCID). We also about several packets to make sure the sender receives
added an optional ACK based reliability layer on top of the information though some ACKs may be lost.
DCCP connection, similar to TCPs reliability scheme. The A weighted average RTT (0.75*RTT +0.25*current RTT)
new CCID profile defines when acknowledgments are sent is calculated using the timestamp echo contained in the
and how to identify the true reasons of packet loss. ACKs. The congestion window (cwnd) size is adjusted
Additional ECN support and ELFN support is used to accordingly reseting the cwnd every RTT based on the
provide network-detected information to the sender. average RTTs collected and keep the same cwnd for this
To implement reliable transmission based on DCCP and RTT as in FAST TCP. This management of cwnd is similar
provide a comparable level of reliability as TCP does, we to the approach in TCP. A timeout timer (RTO) is set for the
added the following functions to DCCP: Buffering of transmitted packets. In Congestion State, the sender sends
received packets at the receivers, retransmission of lost or every RTO, which probed packets with only headers until
an ACK is received. Upon successfully receiving an ACK,

2009 Chinese Control and Decision Conference (CCDC 2009) 4735


the sender resets the RTT and baseRTT, sets the cwnd size OUT flows. While TCP shows severe unfairness when the
to cwnd*Old BaseRTT/baseRTT, and enters the Normal two flows are mixed (see Table 1).
State again.
Table 1. The performance for 2 flows Fairness of TCP, DCCP
and Extended DCCP.
4 EXPERIMENT AND SIMULATION
The test scenario used in the simulation is shown in Figure
3 (inspired by [27]). In the simulation, data is sent from the
wired node to the wireless nodes (1 and 4) that are two hops
away from the access point; or from those wireless nodes to
the wired node via the access point. All wireless nodes are
stationary in the simulation. All data flows are 10MB FTP
flows. We adopt the fairness index fi [28] to evaluate the
inter-session fairness between multiple sessions of the
same bottleneck. fi is defined as follows: if there are N 5 CONCLUSIONS AND FUTURE WORK
concurrent connections in the network and the throughput
achieved by connection Kth is equal to R(K). As is shown in TCP was designed for wired networks, and has benefited
(1) from substantial research efforts over the years. Yet it
shows severe inter-flow fairness challenges over multi-hop
wireless networks, as shown in Section 2. Section 2 reviews
a number of proposals to enhance TCP, with some of the
proposed protocols showing quite promising results.
(1) However, none of these improvements will benefit UDP
Assume that the demand of the flow throughput is streams that are often used in streaming media content.
infinite. We say max-min fairness is achieved when fi Section 3 gives a high-level overview of a
approaches unity, i.e., the throughputs of all sessions are congestion-control approach based on DCCP that could be
equal. The fairness index lies between 0 and 1, 1 being the beneficial to both unreliable data streams and reliable data
fairest situation where every flow gets equal throughput. transfers. Simulation results in NS2 confirm that the
approach improves fairness, providing users with fairness
in multi-hop wireless access network.
Our contribution is to provide a theoretical foundation for
the study of fairness in sensor networks from the max-min
perspective. The proposed algorithm may be used in a
centralized scheme that collects information from the
network, computes max-min rates, and then disseminates
Fig 3. Wireless sensor network simulated scenario.
the rates to the sensors. We also implemented this
congestion control algorithm in NS2. Simulation results
In a set of experiments, two flows traversing a chain of
show improvements on fairness achieved by using
nodes (4 hops) were tested using TCP Reno. Both DCCP
Extended DCCP congestion control algorithm.
and TCP are end-to-end sliding window protocols. Data
packets are transmitted in both directions: packets are sent The work presented here will be further extended in the
from the senders to the receivers and acknowledgements following areas to verify and improve the design. We will
are sent from the receivers to the senders. Senders are conduct more performance runs to verify the test results
allowed to send a window of packets before receiving the under multiple-flow scenarios. We will study the impact of
acknowledgment. This window starts at a constant size and additional loss scenarios caused by broken links during an
is later controlled by the congestion control algorithms active flow on fairness. We will also study and improve
implemented in the protocols. Acknowledgments are valid fairness when mixed Extended DCCP and TCP flows
when sequence numbers of acknowledged packets are co-exist. The core congestion control protocol can be
within the range of the current window. further optimized by tuning the rate control formula and
retransmission timer to optimize the packet sending rate
Each flow could be either an IN flow or an OUT flow, and
and adding new features to the implementation in the
we also varied the RTT for mixed flow scenarios. To
simulation such as support of ECN, to provide additional
evaluate various alternatives, we measured the fairness for
information for the sender to identify network condition
each flow. Table 1 shows the respective Fairness Index
and to adjust the sending rate accordingly.
results for TCP, DCCP and Extended DCCP as transport
protocol under 2 flows throughput. Finally, we are very interested in implementing the
proposed congestion control protocol in wireless sensor
Overall, the simulation results show that DCCP flows have
network test bed to verify the simulation results.
good inter-flow fairness due to the modified congestion
control algorithm, which uses a rate based window control ACKNOWLEDGMENT
algorithm based on the feedback from the
acknowledgments. Compared with TCPs congestion Here, we express the thanks for the profession who have
control, the Extend congestion control algorithm shows given the supports and suggestions to this paper, especially
improved inter-flow fairness for all combinations of IN and the class of students and teachers were led by Professor

4736 2009 Chinese Control and Decision Conference (CCDC 2009)


NIAN Xiao-Hong and Professor LU Wu-Yi. Finally, we [13] J.M. Faffe, Bottleneck Flow Control, IEEE Trans. Comm.,
would also like to thank the teacher YANG Sheng-Yue. All vol. 29, no. 7, pp. 954-962, July 1981.
of them are teacher in School of Information Science and [14] F. Bonomi and K. Fendick, The Rate-Based Flow Control
Engineering, Central South University. Framework for the Available Bit Rate ATM Service, IEEE
Network, vol. 9, no. 2, pp. 25-39, Mar./Apr. 1995.
REFERENCES [15] L. Georgiadis, P. Georgatsos, K. Floros, and S. Sartzetakis,
Lexicographically Optimal Balanced Networks,
[1] Akyildiz, I., Su, W., Sankarasubramaniam, Y., Cayirci, E., IEEE/ACM Trans. Networking, vol. 10, no. 6, 2002.
A Survey on Sensor Networks, IEEE Communications
[16] Dong Suwon. Analysis of Impact on Random Packet Losses
Magazine, Aug. 2002, pp. 102-109.
in Congestion Control: Explicit Form, Proc. IEEE
[2] E. Biagioni and K. Bridges, The Applications of Remote INFOCOM 2007.
Sensor Technology to Assist the Recovery of Rare and
[17] Q. Ma, P. Steenkiste, and H. Zhang, Routing
Endangered Species, Intl J. High Performance Computing
High-Bandwidth Traffic in Max-Min Fair Share Networks,
Applications, Apr. 2003.
Proc. ACM Special Interest Group Data Comm.
[3] H.T. Kung and D. Vlah, Efficient Location Tracking Using (SIGCOMM 96), Aug. 1996.
Sensor Networks, Proc. IEEE Wireless Comm. and
[18] T. Nandagopal, T. Kim, X. Gao, and V. Bharghavan,
Networking Conf. (WCNC 03), Mar. 2003.
Achieving MAC Layer Fairness in Wireless Packet
[4] Gracanin, D., Adams, K., Eltoweissy, M., Data Replication Networks, Proc. MobiCom 00, Aug. 2000.
in Collaborative Sensor Network Systems, in Proc. 25th
[19] H. Luo, J. Cheng, and S. Lu, Self-Coordinating Localized
IEEE International Performance, Computing, and
Fair Queueing in Wireless Ad Hoc Networks, IEEE Trans.
Communications Conference (IPCCC 2006), 10-12 April
Mobile Computing, vol. 3, no. 1, Jan.-Mar. 2004.
2006, pp. 389-396.
[20] Y. Yi and S. Shakkottai, Hop-by-Hop Congestion Control
[5] Md, O, R., Muhammad, M, M., Choong, S, H. A QoS
over a Wireless MultiHop Network, Proc. INFOCOM 04,
Adaptive Congestion Control in Wireless Sensor Network,
Mar. 2004.
The 10th International Conference on Advanced
Communication Technology (ICACT 2008), pp.941-946. [21] K Miller, T Harks,.Utility Max-Min Fair Congestion
Phoenix Park, Republic of Korea, Feb 2008. Control with Time-Varying Delays. Proc. IEEE INFOCOM
2008.
[6] W.R. Heinzelman, J. Kulik, and H. Balakrishnan, Adaptive
Protocols for Information Dissemination in Wireless Sensor [22] A. Sridharan and B. Krishnamachari, Max-Min Fair
Networks, Proc. MobiCom 99, Aug. 1999. Collision-Free Scheduling for Wireless Sensor Networks,
Proc. Workshop Multihop Wireless Networks (MWN 04),
[7] C.-Y. Wan, S.B. Eisenman, and A.T. Campbell, CODA:
Apr. 2004.
Congestion Detection and Avoidance in Sensor Networks,
Proc. ACM Conf. Embedded Networked Sensor Systems [23] V. Gambiroza, B. Sadeghi, and E.W. Knightly, End-to-End
(SenSys 03), Nov. 2003. Performance and Fairness in Multihop Wireless Backhaul
Networks, Proc. MobiCom 04, Sept.-Oct. 2004.
[8] Y. Sankarasubramaniam, O. Akan, and I. Akyildiz, ESRT:
Eventto-Sink Reliable Transport in Wireless Sensor [24] L. Yang et al., Improving Fairness among TCP Flows
Networks, Proc. MobiHoc 03, June 2003. crossing Wireless Ad Hoc and Wired Networks, Proc. of the
4th ACM Int. Symposium on Mobile Ad Hoc Networking &
[9] C. Intanagonwiwat, R. Govindan, D. Estrin, J. Heidemann,
Computing, pp. 5763, Annapolis, USA, 2003.
and F. Silva, Directed Diffusion for Wireless Sensor
Networking, ACM/IEEE Trans. Networking, vol. 11, no. 1, [25] B. Radunovic and J. LeBoudec, Rate Performance
pp. 2-16, Feb. 2002. Objectives of MultiHop Wireless Networks, Proc.
INFOCOM 04, Mar. 2004.
[10] T. He, J.A. Stankovic, C. Lu, and T.F. Abdelzaher, SPEED:
A Stateless Protocol for Real-Time Communication in [26] Kohler E., Handley M., Floyd S., Datagram Congestion
Sensor Networks, Proc. Intl Conf. Distributed Computing Control Protocol (DCCP), RFC 4340, IETF, March 2006.
Systems (ICDCS 03), May 2003. [27] C. Jin, D. Wei, S. H. Low, FAST TCP: Motivation,
[11] S. Pilisof, R. Ramjee, D. Raz, Understanding TCP Fairness Architecture, Algorithms, Performance, Proc. of the 23rd
Over Wireless LAN, Proc. of the 22nd Annual Joint Conf. of Conf. of the IEEE Communication Society, pp. 8-94, Hong
IEEE Computer and Communications Societies, pp. 863872, Kong, China, March 2004.
April 2003. [28] R. Jain, The Art of Computer Systems Performance Analysis:
[12] K. Xu et al., TCP Behavior across Multihop Wireless Techniques for Experimental Design, Measurement,
Networks and the Wired Internet, Proc. of the 5th Int. Simulation, and Modeling. John Wiley and Sons, 1991.
Workshop on Wireless Mobile Multimedia, pp. 207218,
Seattle, USA, Sept. 2002.

2009 Chinese Control and Decision Conference (CCDC 2009) 4737

Das könnte Ihnen auch gefallen