Enhanced Backoff Scheme in CSMA/CA for IEEE 802.

11
Wen-Kuang Kuo and C.-C. Jay Kuo
Department of Electrical Engineering-Systems and Integrated Media Systems Center University of Southern California, Los Angeles, CA 90089-2564 Abstract-- An enhanced CSMA/CA (Carrier Sense Multiple Access with Collision Avoidance) protocol to be used in the Medium Access Control (MAC) layer of the IEEE 802.11 standard for wireless local area networks (wireless LANs) is proposed in this work. In wireless LANs, the CSMA/CA protocol supports asynchronous data transfer, and adopts an acknowledgment mechanism to confirm successful transmissions and a handshaking mechanism to reduce collisions. In both cases, a binary exponential backoff mechanism is used. The enhanced protocol improves the exponential backoff scheme by dynamically adjusting the contention window (CW) around the optimal value. Moreover, an analytical model based on the Markov chain is developed to analyze the system performance in terms of throughput and delay. Numerical results are presented to show the effect of the proposed backoff mechanism. I. INTRODUCTION Since wireless local area networks (wireless LANs) can provide more flexible and convenient connections than wired networks, there is an increasing demand for wireless LANs recently. To speed up the design of wireless LANs, the IEEE 802.11 study group proposes an international standard [1] for wireless LANs. The standard defines detail functions for both the Medium Access Control layer and the Physical Layer. In the IEEE 802.11 standard, the distributed coordination function (DCF) is defined for asynchronous data transmissions while the optional point coordination function (PCF) is used to support time-bounded data transfer. The DCF shares access to the medium based on the CSMA/CA protocol. Note that collision detection is not used in wireless LANs, since a station is unable to detect the channel and transmit data simultaneously. To address this issue, robust CSMA/CA was developed instead. The basic idea of CSMA/CA is listen before talk, where a station which desires to transmit must probe the medium before transmission to determine whether another user is transmitting. CSMA/CA employs an immediate positive acknowledgment scheme to make sure successful reception of packets. The receiving station sends the acknowledgment packet after a short time interval. If an acknowledgment is not received, the packet is considered lost and a retransmission is arranged. There are some simulation and analytical studies on the performance of the CSMA/CA protocol. In [2], the theoretical throughput limit of wireless LANs according to a p-persistent mechanism was analyzed. However, the work in [2] does not consider the effect of the contention window (CW) and the binary exponential backoff scheme used by IEEE 802.11. In [3], the Markov chain was used to analyze the saturated throughput of IEEE 802.11. In this work, we modify the scheme proposed in [3] by taking into account the CW resetting scheme under busy medium conditions. We show how the modified scheme affects the behavior of the backoff mechanism. Furthermore, a Markov chain model is proposed to analyze the throughput and delay performance of the new scheme. Finally, the derived analytical results are verified by computer simulation. This paper is organized as follows. In Section 2, we briefly review both the basic and the RTS/CTS access mechanisms of the DCF and the IEEE 802.11 MAC layer. Section 3 introduces the modified contention window adjustment scheme, the Markov model for performance evaluation, and the saturation throughput and access delay analysis. Section 4 validates theoretical derivations by simulation, which is followed by some discussion on the IEEE 802.11 MAC performance. Finally, some concluding remarks are given in Section 5.
II. IEEE 802.11 MEDIUM ACCESS CONTROL (MAC) LAYER AND
DISTRIBUTED COORDINATION FUNCTION (DCF)

The IEEE 802.11 standard defines both the medium access control (MAC) and the physical (PHY) layers. The reference model of IEEE 802.11 is shown in Fig.1. In this work, we focus on the discussion and enhancing the capability of the MAC layer. One of the most important functions of the MAC layer of IEEE 802.11 is to coordinate the wireless medium access procedure. The fundamental access method in the IEEE 802.11 MAC protocol is the distributed coordination function (DCF), which is based on the CSMA/CA MAC protocol. In addition to the DCF, the IEEE 802.11 standard also provides an alternative access method, called the point coordination function (PCF), which is an access method similar to a polling system and uses a point coordinator to determine which station to transmit. The DCF consists of the basic access mode as well as the optional RTS/CTS access mode. These two access methods are described below. In basic access mode, a station has to sense the channel to determine whether another station is transmitting before initiating a transmission. If the medium is sensed to be free

0-7803-7955-1/03/$17.00 (C) 2003 IEEE

for a DIFS interval, the transmission may proceed. On the other hand, if the medium is busy, the station must defer its transmission until the end of the current transmission. Then, it will wait for an additional DIFS interval and generate a random backoff delay to initialize the backoff timer before transmission. The backoff timer is decreased as long as the medium is sensed as idle and frozen when a transmission is detected on the channel, and resumed when the medium is sensed as idle again for more than a DIFS interval. Only when the backoff timer reaches zero, the station transmits its packet. For IEEE 802.11, time is slotted in a basic time unit, denoted by Slot_Time, which is equal to the time needed to detect the transmission of a packet from any other station. The backoff delay is an integer uniformly chosen in the range [0, W-1], which is called the backoff window or the contention window, and corresponds to the number of time slots that must elapse before the station can sense the channel again. The initial backoff window size is set to W=CWmin + 1. If two or more stations decrease their backoff timer to zero at the same time, collisions occur. At this situation, the backoff window size is doubled for each retransmission until it reaches the maximum value, W=CWmax + 1. Note that IEEE 802.11 employs the inter-frame space (IFS) time to control the priority access to the channel between packet transmissions. Three IFS intervals are defined: the short IFS (SIFS), the point coordination function IFS (PIFS), and the DCF-IFS (DIFS). SIFS is the smallest while DIFS is the largest. Stations that are required to wait a SIFS interval have a higher priority to access the medium than those that are required to wait a PIFS or a DIFS interval. When receiving a packet correctly, the destination station waits for a SIFS interval immediately after the reception of this packet and transmits an acknowledgement back to the source station, confirming the correct reception. If the source station does not receive an acknowledgement due to collisions or transmission errors, it reactivates the backoff algorithm after the channel remains idle for at least an extended IFS (EIFS) interval. In IEEE 802.11, a station senses the wireless medium via physical carrier sensing, which detects the activity in the air interface through the relative signal strength of other stations, and virtual carrier sensing, which is performed by the MAC layer. To incorporate the virtual carrier sensing mechanism, a station has to send the MPDU (MAC protocol data unit) duration information in the header of request-to-send (RTS), clear-to-send (CTS) and data frames. The MPDU duration information indicates the amount of time the wireless medium is to be reserved for transmitting the data and returning ACK frames after the end of the current frame. The station in the same basic service set (BSS) uses this information to update its network allocation vector (NAV) that represents the amount of time it has to defer in accessing the wireless medium. By using the virtual carrier sensing mechanism, stations within the same BSS learn how long the channel will be used for this data transmission. In fact, the

virtual carrier sensing, which is especially important in a wireless environment, is an enhancement of the physical carrier sensing. In RTS/CTS access mode, after obtaining the right to access the channel, the source station sends a RTS frame prior to data transmission to announce the upcoming transmission. When the destination station receives the RTS frame, it will transmit a CTS frame after a SIFS interval. Both the RTS and CTS frames are short control frames. The source station is allowed to transmit its packets only if it receives the CTS frame correctly. Fig. 3 illustrates the operation of this mechanism. Because a station cannot transmit and listen to the channel at the same time, it will continue transmitting the whole MPDU even if a collision occurs. If the MPDU is large, much of the channel bandwidth is wasted. Therefore, the advantage of exchange of short RTS/CTS frames before data transmission is to avoid long collisions and reduce the amount of bandwidth wasted in collisions. III. PROPOSED BACKOFF SCHEME As described in Section 2, a binary exponential backoff mechanism is used with CSMA/CA. Whenever a backoff occurs, the backoff time is uniformly chosen in the range (0, CW-1). After each unsuccessful transmission, the backoff window size is doubled, up to a maximum value 2mW, where W equals to (CWmin+1) and 2mW equals to (CWmax+1). Once the backoff window size reaches CWmax, it will stay at the value of CWmax until it is reset. The value of w will be reset to CWmin after every successful transmission of a data frame or a RTS frame, or when a retry counter reaches its limit. Since a station uses CW to control the backoff counter for data transfer, how to set CW will affect the performance of the CSMA/CA protocol. To optimize the performance of the CSMA/CA protocol, we consider the backoff procedure as a progress to search the optimal value of CW. The problem is that the CW resetting scheme in current IEEE 802.11 standard does not obey this optimization progress. Therefore, the CSMA/CA protocol is difficult to achieve the optimal performance. To solve this problem, we propose a dynamic CW resetting scheme to let the backoff counter oscillate around the optimal value. Our method can be easily implemented in the current IEEE 802.11 standard without complex calculation and runtime estimation as done in [2]. The proposed scheme is described as follows: A. After a successful transmission, w is set to the value max[w/2, CWmin +1]. B. Whenever a transmission fails, w is set to the value min[2w, CWmax +1].

0-7803-7955-1/03/$17.00 (C) 2003 IEEE

The second major contribution of this work is mathematical evaluation of the throughput and delay performance of the proposed backoff scheme by considering busy channel conditions. Note that we do not adopt the p-persistent scheme as described in [2], since it is not consistent with the binary exponential backoff scheme adopted by IEEE 802.11. Furthermore, a new Markov chain model is proposed to analyze the effect of the new backoff scheme. The analysis includes two parts: (1) the behavior of a station is examined using the Markov chain, and (2) the throughput and delay of both basic and RTS/CTS access modes are derived. To facilitate the analysis, we make the following assumptions. The contending stations are assumed to be a fixed number, denoted by n. The function b(t) is defined to be the stochastic process representing the random backoff window value for a given station at slot time t while s(t) is the stochastic process representing the backoff stage i for the same station at time t. Thus, the two-dimensional process, {s(t), b(t)}, is a discrete-time Markov chain. Furthermore, it is clear from the backoff algorithm described above that the probability p for transmitted packets to collide as well as the probability pb for the channel to be in a busy status are independent of the backoff procedure. The discrete-time Markov chain model corresponding to the problem described above is shown in Fig. 4. We use the similar mathematical operation in [3] and [5] to analyze the proposed Markov chain model.
IV. MODEL VALIDUATION AND DISCUSSION

is the number of stations and Y-axis is the saturation throughput. Here, m stands for the number of the backoff stage and w stands for the initial backoff window size. In Figs.2 and 3, we plot results with the same number of the backoff stage but a different initial backoff window size for each scheme. We have the following observations in these two figures. First, we see that the analytical model is accurate in the sense that it is consistent with the general trend of simulation results. Second, the proposed scheme has a higher throughput since it can decrease the chance of collision by adaptively adjusting the contention window. Third, the saturation throughput depends on the initial backoff window size. When the number of stations increases, the saturation throughput of the basic mode decreases since the probability of collision becomes larger. For the RTS-CTS mode, saturation throughput degradation is less significant when the number of stations increases. Besides, the basic access mode has a higher throughput when we choose a larger initial backoff window size. However, for the RTS-CTS access mode, the throughput is less improved even with a larger initial backoff window size. This can be explained by the fact that a larger backoff window can decrease the probability of collision and the number of retransmission for the basic access mode. In contrast, the RTS/CTS access mode by itself can avoid long collision and the associated waste of the bandwidth when collision occurs. Figs. 4 and 5 depict the dependence of the WLAN throughput on the number of the backoff stages m. In both figures, the X-axis is the number of stations and the Y-axis is saturation throughput. In both figures, the initial backoff window sizes are the same but the number of the backoff stage varies. From these two figures, we show the effectiveness of the proposed scheme. To increase the backoff stage will enlarge the contention window and reduce collisions. We observe that the basic mode improves its throughput when a larger number of stations are active. For the RTS/CTS modes, the improvement is less obvious. Figs. 6 and 7 show the access delay as a result of a different initial backoff window size but with same number of backoff stages for both access modes. In both figures, the X-axis is the number of stations and the Y-axis is saturation delay in msec. These two figures demonstrate that the proposed scheme has a lower access delay. This is due to the fact that the number of retransmissions is reduced in the proposed scheme. We also observe that the RTS/CTS mode has a lower access delay since a smaller amount of overhead is needed for transmission. When the number of stations attempting to transmit increases, the number of collisions also increases. Hence, the access delay becomes longer. Figs. 8 and 9 depict the relationship between the access delay and a different number of backoff stages with the same initial backoff window size. In both figures, the X-axis is the number of stations and the Y-axis is saturation delay in msec. These two figures say that the WLAN protocol suffers a higher delay when the backoff procedure includes more

To validate the analytic model derived in Sec. 3, we compare analytical results with simulation results conducted with OPNET [8]. To demonstrate the advantage of the proposed scheme, we compare its performance, including the saturation throughput and access delay, with that of the original backoff scheme in IEEE 802.11. To gain deeper insights into the IEEE 802.11 DCF performance, extensive analytical and simulation results are shown and discussed. Parameters used in our analysis and simulation are summarized in Table 1.
Packet payload 1024 bytes MAC header 34 bytes PHY header 16 bytes ACK 14 bytes+PHY header RTS 20 bytes+PHY header CTS 14 bytes+PHY header Channel bit rate 1 M bps Propagation delay 1 s Slot time 20 s SIFS 28 s DIFS 128 s ACK_timeout 200 s CTS_timeout 200 s Table 1: System parameters used in analysis and simulation.

Figs. 2 and 3 show the saturation throughput of the RTSCTS and the basic access modes. In both figures, the X-axis

0-7803-7955-1/03/$17.00 (C) 2003 IEEE

backoff stages. When the number of collisions increases, the stations will choose a higher backoff statge (i.e. a larger backoff window size), which increases access delay.
V. CONCLUSION

(1-P)/W 0 0, 0 1-Pb 0, 1 Pb 1-Pb 1-Pb 0, W 0-2 Pb 1-Pb 0, W 0-1 Pb

P/W1 1-Pb P/W i

In this research, we proposed an enhanced backoff scheme for the IEEE 802.11 DCF protocol, which is effective yet easily implemented in the current standard. A new Markov model was presented to compute the theoretical saturation throughput and access delay for both the basic and the RTS/CTS access mechanisms. Based on the analytic model, we demonstrated the performance enhancement of the proposed contention window selection scheme. Besides, the relationship between the performance improvement, the number of backoff stages and the initial backoff window size was investigated. We observe that the performance of the basic access mode strongly depends on these parameters while that of the RTS/CTS access mode is less affected by varying these parameters. Finally, the RTS/CTS access mode has a higher throughput and a lower access delay than the basic access mode under the saturated traffic condition.
ACKNOWLEDGEMENTS

i-1, 0

i-1, 1 Pb (1-P)/Wi-1

1-Pb

1-Pb

i-1,W i-1-2

1-Pb

i-1,W i-1-1

Pb

Pb

i, 0

1-Pb P/Wi+1

i, 1 Pb

1-Pb

1-Pb

i, W i-2 Pb

1-Pb

i, Wi-1 Pb

(1-P)/Wm-1 m, 0 1-Pb m, 1 Pb 1-Pb 1-Pb m,W m-2 Pb 1-Pb m,Wm-1 Pb

P/Wm

Figure 1: The Markov chain model of the proposed scheme.

0.87 0.85 0.83 0.81 0.79 0.77 0.75 1 5 9 13 17 21 25 29 33 37 41 45 49

The research has been funded by the Integrated Media Systems Center, a National Science Foundation Engineering Research Center, Cooperative Agreement No. EEC-9529152. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of the National Science Foundation.
REFERENCES

w=32 proposed theoretical w=32 original theoretica w=16 proposed theoretica w=16 original theoretica w=32 proposed simulation w=32 original simulation w=16 proposed simulation w=16 original simulation

1. 2.

3. 4. 5.

IEEE standard for Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications, ISO/IEC 8802-11:1999(E), Aug. 1999 F.Cali, M.Conti, E.Gregori. Dynamic Tuning of the IEEE 802.11 Protocol to Achieve a Theoretical Throughput Limit, IEEE/ACM Trans. On Networking, V8, N6, pp.785-799, Dec. 2000 G. Bianchi, Performance Analysis of the IEEE 802.11 Distributed Coordination Function, V18, N3, pp.53557-47, March 2000 B.P. Crow, J.G. Kim, IEEE 802.11 Wireless Local Area Networks, IEEE Communications magazine, Sept. 1997 E. Ziouva , T. Antonakopoulos, CSMA/CA Performance under High Traffic Conditions: Throughput and delay Analysis Computer communications, V 25, pp. 313-321, 2002

Fig. 2 Saturation throughput of the RTS-CTS access mode with m=3

0.82 0.8 0.78 0.76 0.74 0.72 0.7 0.68 0.66 0.64 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49
w=32 proposed simulation w=32 original simulation w=16 proposed simulation w=16 original simulation w=32 original theoretical w=16 proposed theoretical w=16 proposed theoretical w=32 proposed theoretical

Fig. 3 Saturation throughput of the basic access mode with m=3.

0-7803-7955-1/03/$17.00 (C) 2003 IEEE

0.87 0.85 0.83 0.81 0.79 0.77 0.75 1 5 9 13 17 21 25 29 33 37 41 45 49

m =5 proposed theoretical m =3 proposed theoretical m =5 original theoretica m =3 original theoretica m =5 proposed sim ulation m =3 proposed sim ulation m =5 original sim ulation m =3 originall sim ulation

800 700 600 500 400 300 200 100 0 1 5 9 13 17 21 25 29 33 37 41 45 49

m = 5 proposed theoretical m = 3 proposed theoretical m = 5 original theoretical m = 3 original theoretical m = 5 proposed sim ulation m = 3 proposed sim ulation m = 5 original sim ulation m = 3 original sim ulation

Fig.4 Saturation throughput of the RTS-CTS access modes with w=16

Fig.8 Saturation delay of the RTS-CTS access mode with w=16.

900 800 700 600 500 400 300 200 100 0 1 5 9 13 17 21 25 29 33 37 41 45 49
m=5 proposed theoretical m=3 proposed theoretical m=5 original theoretical m=3 original theoretical m=5 proposed simulation m=3 proposed simulation m=5 original simulation m=3 original simulation

0.82 0.8 0.78 0.76 0.74 0.72 0.7 0.68 0.66 0.64 1 5 9 13 17 21 25 29 33 37 41 45 49

Fig. 5 Saturation throughput of the basic access mode with w=16

m=5 proposed theoretical m=3 proposed theoretical m=5 original theoretical m=3 original theoretical m=5 proposed simulation m=3 proposedl simulation m=5 original simulation m=3 original simulation

Fig.9 Saturation delay of the basic access mode with w=16.

500 400 300 200 100 0 1 5 9 13 17 21 25 29 33 37 41 45 49

w=32 proposed theoretical w=32 original theoretical w=16 proposed theoretical w=16 original theoretical w=32 proposed simulation w=32 original simulation w=16 proposed simulation w=16 original simulation

Fig.6 Saturation delay of the RTS-CTS access mode with m=3.

w=32 proposed theoretical w=32 original theoretical w=16 proposed theoretical w=16 original theoretical w=32 proposed simulation w=32 original simulation w=16 proposed simulation w=16 original simulation

600 500 400 300 200 100 0 1 5 9 13 17 21 25 29 33 37 41 45 49

Fig.7 Saturation delay of the basic access mode with m=3.

0-7803-7955-1/03/$17.00 (C) 2003 IEEE