International Journal of Emerging Technology and Advanced Engineering Website: (ISSN , Volume 2, Issue 1, January 2012)

Transcription

1 Performance Improvement of MAC by enhancements in DCF Vikram Jain 1, Siddharth Dutt Choubey 2, Rohit Singh 3, Sandeep Gupta 4 1 Research Scholar, Singhania University, INDIA 2,3,4 ME / MTech (CS Dept.) SRIT, Jabalpur, INDIA 1 vikram.srit@gmail.com 2 siddharth.choubey@gmail.com 3 rohit.singhthakur2@gmail.com 4 sandeepgupta_mca@rediffmail.com Abstract--In the protocol, the fundamental mechanism to access the medium is called distributed coordination function (DCF). This is a random access scheme, based on the carrier sense multiple access with collision avoidance (CSMA/CA) protocol. Retransmission of collided packets is managed according to binary exponential backoff (BEB) rules. BEB increases the waiting time of each node exponentially by 2 with every unsuccessful transmission. Also, after each successful transmission, the backoff stage will resume to the initial stage 0, and the contention window will be set to minimum regardless of network conditions such as number of competing nodes. This method, referred to as heavy decrease tends to work well when the number of competing nodes is less. When the number of competing nodes increases, it will be shown to be ineffective, since the new collisions can potentially occur and cause significant performance degradation. In this paper, we propose some modifications in the basic BEB algorithm, and propose an enhanced DCF(EDCF) and evaluate its performance through simulations. Also we propose some additional modifications in the basic DCF to enhance the performance of Medium Access Control Protocol. Keywords MAC; DCF; CSMA/CD;BEB;WLAN I. INTRODUCTION WLANs have great potential for both military and commercial applications. In a WLAN, nodes transmit packets in an unsynchronized fashion. The protocol employed in the medium access control (MAC) layer is responsible for coordinating access to the shared channel while minimizing conflicts. Hence it is important to design an efficient and effective MAC protocol. The fundamental access method of the IEEE MAC is a DCF known as carrier sense multiple access with collision avoidance (CSMA/CA). The DCF must be implemented in all stations, for use within both ad-hoc and infrastructure network configurations. For a station to transmit, it shall sense the medium to determine if another station is transmitting. If the medium is not determined to be busy, the transmission may proceed. The CSMA/CA distributed algorithm mandates that a gap of a minimum specified duration exist between contiguous frame sequences. A transmitting station must ensure that the medium is idle for this required duration before attempting to transmit. If the medium is determined to be busy, the station shall defer until the end of the current transmission. After deferral, or prior to attempting to transmit again immediately after a successful transmission, the station shall select a random backoff interval and should decrement the backoff interval counter while the medium is idle. II. OPERATION MODE OF CONVENTIONAL DCF In , the DCF is the fundamental access method used to support asynchronous data transfer on a best effort basis. As specified in the standards, the DCF must be supported by all the stations in a basic service set(bss). The DCF is based on Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA). CSMA/CD is not used because a station is unable to listen to the channel while transmitting. In CS is performed both at physical layer, which is also referred to as physical carrier sensing, and at the MAC layer, which is known as virtual carrier sensing. The PCF in the is a polling-based protocol, which is designed to support collision free and real time services. There are two techniques used for packet transmitting in DCF. The default one is a two-way handshaking mechanism, also known as basic access method. A positive MAC acknowledgement (ACK) is transmitted by the destination station to signal the successful packet transmission. The other optional one is a four-way handshaking mechanism, which uses RTS/CTS technique to reserve the channel before data transmission. 337

2 Before transmitting a packet, a station operating in RTS/CTS mode reserves the channel by sending a special Request-To-Send short frame. The destination station acknowledges the receipt of an RTS frame by sending back a Clear-To-Send frame, after which normal packet transmission and ACK response occurs. Since collision may occur only on the RTS frame, and it is detected by the lack of CTS response, the RTS/CTS mechanism allows to increase the system performance by reducing the duration of a collision when long messages are transmitted. As an important side effect, the RTS/CTS scheme designed in the protocol is suited to combat the so-called problem of Hidden Terminals, which occurs when pairs of mobile stations result to be unable to hear each other. However, the drawback of RTS/CTS mechanism is increased overhead for short data frames. A station with a new packet to transmit monitors the channel activity. If the channel is idle for a period of time equal to a distributed inter-frame space (DIFS), the station transmits. Otherwise, if the channel is sensed busy (either immediately or during the DIFS), the station persists to monitor the channel until it is measured idle for a DIFS. At this point, the station generates a random backoff interval before transmitting (this is the Collision Avoidance feature of the protocol), to minimize the probability of collision with packets being transmitted by other stations. In addition, to avoid channel capture, a station must wait a random backoff time between two consecutive new packet transmissions, even if the medium is sensed idle in the DIFS time. For efficiency reasons, DCF employs a discrete-time backoff scale. The time immediately following an idle DIFS is slotted, and a station is allowed to transmit only at the beginning of each slot time. The slot time size is set equal to the time needed at any station to detect the transmission of a packet from any other station. As shown in Table I, it depends on the physical layer, and it accounts for the propagation delay, for the time needed to switch from the receiving to the transmitting state, and for the time to signal to the MAC layer the state of the channel (busy detect time). A. Binary Exponential Backoff Algorithm DCF adopts an exponential backoff scheme. At each packet transmission, the backoff time is uniformly chosen in the range (0, W-1). The value W is called contention window, and depends on the number of transmissions failed for the packet. At the first transmission attempt, is set equal to a value CWmin, called minimum contention window. After each unsuccessful transmission, W is doubled, up to a maximum value CWmax = 2^m * CWmin. 338 The values and reported in the final version of the standard are summarized in Table 1. PHY Slot Time FHSS 50 µsec DSSS 20 µsec IR 8 µsec CWmin CWmax Table 1 : Slot Time, Minimum and Maximum Contention Window Values for three PHY specified by : Frequency Hopping Spread Spectrum (FHSS), Direct Sequence Spread Spectrum (DSSS) and InfraRed (IR) The backoff time counter is decremented as long as the channel is sensed idle, frozen when a transmission is detected on the channel, and reactivated when the channel is sensed idle again for more than a DIFS. The station transmits when the backoff time reaches zero. B. Problems With Existing Architecture From the above discussion we can see that DCF resolves collision through Contention Window and backoff time. As specified in the original standard, after each successful transmission, the backoff stage will resume to the initial stage 0, and the contention window will be set to CWmin regardless of network conditions such as number of competing nodes. This method, referred to as heavy decrease tends to work well when the number of competing nodes is less. When the number of competing nodes increases, it will be shown to be ineffective, since the new collisions can potentially occur and cause significant performance degradation. The operation of existing DCF protocol can be summarized from the following figure

3 For example, let us assume that the current backoff stage is i with contention window CW( i ) = 2^i * CWmin, and there is a successful transmission, the next backoff stage will be stage 0 with contention window CW( 0 ) = 31 according to the specification. But if the number of competing nodes is large enough (>>31), the new collision will likely occur at the backoff stage 0. The main argument is that since the current backoff stage is i some collision must have occurred recently at the previous stage. Now if the number of current competing nodes is larger than or close to CW( i ), and if the backoff stage is set to 0, there is a high probability that new collisions will happen. So resetting the contention window after every successful transmission is an inefficient approach if the number of nodes is large. Thus, the operation of the BEB algorithm can be summarized as follows: The modified backoff algorithm upon unsuccessful transmission can therefore be expressed as, CW = min [b*cw, CWmax], upon collision, where b is the base value (e.g., b = 2 in BEB). Also EDCF attempts to avoid useless collisions through gentle decrease of contention window. The collision resolution stage evolution for EDCF is shown in the following figure: CW = min [2*CW, CWmax], upon collision (1) CW = CWmin, upon success (2 ) We also observe that The BEB algorithm causes a fast build-up (i.e., growth-rate) of waiting times spreading the backlog traffic over a larger time frame. However, this fast build-up of waiting time with increasing number of occurrence of collisions might not be appropriate for a MANET, wherein the contending nodes might leave the geographical location of contention itself after a short while due to their mobility. In view of this, we conjecture that it may not be necessary to make a node wait for a duration that builds up exponentially with a binary base. III. PROPOSED ENHANCED DCF (EDCF) BEB algorithm stated in the standard might introduce long transmission delays without any significant benefit in respect of contention resolution due to node mobility. In fact, contending mobiles are likely to move to a different location during a large waiting time for retransmission after a packet collision and might get involved into another independent collision process elsewhere. Thus, the large growth rate in waiting time might not be appropriate in a mobile scenario as in a WLAN. Bearing this in mind, I have proposed a modification to BEB algorithm suitable for WLANs. In particular, the backoff time is increased exponentially, but with a reduced base value (less than 2) after each unsuccessful transmission until a prescribed maximum value (CWmax) is reached. 339 The difference between EDCF and DCF is that EDCF will halve the contention window value if there is a successful transmission. On the contrary, DCF will reset the contention window once there is successful transmission or the retry count overruns the maximum threshold (CWmax). Also, if the retry count reaches the maximum threshold, and a collision is encountered, the contention window is kept as it is and is not reset as in the conventional DCF protocol. Hence, the modified backoff algorithm upon successful transmission can therefore be expressed as, CW = min [CWmin, (CWmax )/2], upon transmission. successful Hence the complete BEB algorithm for EDCF can be expressed as: CW = min [b*cw, CWmax], upon collision where b is the base value (e.g., b = 2 in BEB). CW = max [CWmin, (CWmax )/2], upon transmission. IV. SIMULATION successful Design and implementation of the proposed protocol has been carried out using Global Mobile Information System Simulator (GloMoSim) which is a scalable simulation environment for large wireless and wired communication networks. GloMoSim uses a parallel discrete-event simulation capability provided by Parsec.

4 Our simulation considered a network of 50 mobile nodes placed randomly within a 1000 x 1000 m2 area. Node movement is modeled using the random waypoint mobility model (RWMM), which is widely used in MANET simulations. In RWMM, nodes move at a speed uniformly distributed in [MIN SPEED, MAX SPEED]. Each node begins the simulation by moving towards a randomly chosen destination (waypoint). Whenever a node reaches a waypoint, it rests for a pause time. It then chooses a new waypoint and moves towards the same. This process is repeated until the end of simulation time. In our simulations, however, pause time is set at zero (i.e., nodes move continuously throughout the simulation period). This is done to study the impact of continuous node mobility (i.e., worst-case scenario) on the network performance. Constant bit rate (CBR) data sessions among randomly chosen source-destination pairs (SDPs) are used. For example, with 10 SDPs amongst 50 nodes, 10 source nodes and 10 destination nodes (i.e., 20 nodes in total) will be engaged in data transfer. However, during this data transfer process, all of the 50 nodes (including the above 20 nodes) will operate in the background for providing necessary support (i.e., routing/forwarding) to the ongoing communication process in the network. 4.1 SIMULATION PARAMETERS The protocol was tested for different values of exponential factor ranging from 2 to 2. The performance for each value was analyzed and compared against the conventional DCF. V. RESULTS OBTAINED The new protocol was tested for different backoff values. It was found that the values 1.5, 1.8 and 2 gives a better performance compared to other values. A series of simulations were performed and the new protocol was tested against the conventional protocol for parameters such as Average end-to-end-delay, throughput and average packet loss ratio. The following graphs were plotted from the obtained output A. Throughput Analysis EDCF with backoff factor 1.5 performs better than the conventional DCF protocol when the number of nodes in the network is less. However, as the node density increases, throughput for the original DCF drops significantly. EDCF with backoff value 1.8 gives better values for greater number of nodes compared to other values. The throughput drops for all the values as the node density increases in the network. The following table shows the simulation parameters that were used for simulation Parameter Value Mobility Model Random Waypoint Speed of Mobile Node Uniformly distributed Between [0,10] m/sec Propagation model Two Ray Transmission range 250 meters Of each node Channel Frequency 2.4 GHz Data Rate 2 Mbps Network Protocol IP No of SDPs 5,11,20,25,40 MAC protocol DCF with backoff values 1.5, 1.8, 2 Routing Protocol AODV Table 2 Salient Simulation Parameters Figure 5.1 Throughput Vs Number of nodes B. Average End-to-End Delay The following graph shows the average end-to-end delay for backoff factors 1.5, 1.8 and 2, against the number of nodes in the network. Whenever, there is a successful transmission, contention window will be halved. If the new value of contention window is lesser than the minimum value, the minimum value is kept. 340

5 VI. SOME ADDITIONAL PROPOSED MODIFICATION A. A Gentle Approach Towards decreasing Contention Window. Figure 5.2 Average end to end delay vs Number of nodes The average end-to-end delay does not differ much for the different values of backoff factor when the number of nodes in the network is few. However, as the number of nodes increases in the network (>25), the average end-toend delay increases almost exponentially with the conventional DCF protocol. EDCF with backoff factor 1.5 gives the least average end-to-end delay for high node density network. 5.3 Average packet loss ratio analysis The following graph shows the average packet loss ratio for backoff values 1.5, 1.8 and 2 against the number of nodes in the network. Packet loss ratio is calculated as the ratio number of packets transmitted by the sender to the total number of packets successfully received by the destination. DCF decreases the contention window to its initial value upon every successful transmission, which essentially assumes that network is under low traffic load. But in certain hotspot areas where the network is highly loaded, we can halve the contention window by a factor of 2 upon k successful transmission rather than halving it upon every successful transmission. The main challenge lies in computing the optimal value of k. The intuition is that with many (or few) competing nodes, it requires large(or small) value of k. So if the number of current competing nodes can be obtained, we can intelligently adjust the parameter k. But it is usually expensive to precisely obtain the number of competing nodes in a dynamic and distributed environment where nodes consistently move in and out of network. B. Obtaining the number of Competing Nodes. Although as mentioned, in a distributed environment, it is difficult to find the actual number of currently competing nodes, but, if we can have some kind of Admission Control or Node Registration method, in which a node registers itself in the network before it starts transmitting, we can have a rough estimate of total participating nodes with which we can find the optimal number of k, as mentioned in the previous method. Also we can regulate the value of Contention Window between an optimal minimum and maximum value instead of resetting it and doubling it upon every successful/unsuccessful transmission. Figure 5.3 Packet Success Rate Vs Number of Nodes The above graph suggests that as the number of nodes in the network increases, conventional DCF s packet success rate decreases significantly. In other words, conventional DCF has the maximum average packet loss ratio. EDCF with backoff value 1.5 has the highest success rate and hence the minimum packet loss ratio. 341 C. Providing a QOS We can propose methods to provide service differentiation between high and low priority traffics with focus on fairness. The parameters that we can employ here for differentiation are Inter Frame Spacing(IFS) and CW increasing factor. These two important parameters are used to differentiate between traffics in many researches. In a competition between two nodes which are sensing the medium with different IFS, the node having smaller IFS will win. Therefore, we allocate smaller IFS to the high priority traffics. Therefore the chance of the node with smaller CW to send its packet is higher. We dedicate smaller IFS and CW increasing factor to higher priority traffics.

6 Also, we can have different values of minimum and maximum CW for different traffic types. For example, we can have a contention window ranging from from highest priority traffic, so that nodes involved in sending high priority data gets access of the medium more frequently than the normal priority nodes. VI. CONCLUSION AND FUTURE WORK In this paper, we have proposed a new DCF protocol known as EDCF by modifications in the Binary Exponential Backoff algorithm. Different values for backoff factors were tested and compared against the conventional IEEE DCF protocol. IEEE has several disadvantages in that it s throughput decreases as the number of nodes in the network increases, average end to end delay is more, and there is a higher packet loss ratio in high node density networks. Simulation for the new protocol were carried out using GloMoSim simulator and simulation results shows that EDCF with backoff value 1.5 gives better performance than other values and has several advantages such as 1. It obtains higher throughput than traditional DCF especially when network comprises of large number of competing nodes. 2. It has lower end to end delay and could be deployed in delay sensitive applications. 3. It drops fewer packets in MAC level and can easily be extended to support priority applications. 4. Finally, EDCF is very easy to deploy. It does not need to estimate number of competing nodes in the network and requires no change in the message structure and access procedures of DCF. Also with the suggested improvements, the performance of MAC can further be improved. The new DCF protocol with these suggested improvements and is expected to perform much better and give optimal results. Our future work will include simulation and performance study of proposed work for multimedia traffic & Designing of mechanism for obtaining the number of competing nodes so that QoS can provided. References: [1] IEEE standard, Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, IEEE Std. June [3] Chonggang Wang, Weiwen Tang, Kazem Sohraby, Bo Li, A simple mechanism on MAC layer to improve the performance of IEEE DCF, Broadband Networks, BroadNets Proceedings of First International Conference on Broadband networks 2004, Page(s): [4] Guanghong Wang, Yantai Shu, Liang Zhang, Oliver W.W. Yang, Improving DCF in IEEE Wireless LAN, IEEE CCECE CanadianConference on Electrical and Computer Engineering, Volume 2, 4-7 May 2003 Page(s): [5] Yunli Chen, Qing-An Zeng, Dharma P. Agrawal, Performance analysis and enhancement for IEEE MAC protocol, ICT th International Conference on Telecommunications, 2003, Volume 1, 23 Feb.-1 March 2003 Page(s): [6] Manshaei M.H., Cantieni G.R., Barakat C., Turletti T, Performance analysis of the IEEE MAC and physical layer protocol, Sixth IEEE International Symposium on a World of Wireless Mobile and Multimedia Networks, WoWMoM June 2005 Page(s): [7] Haitao Wu, Yong Peng, Keping Long, Shiduan Cheng, A simple model of IEEE wireless LAN, International Conferences on Info-tech and Info-net, Proceedings. ICII Beijing Volume 2, 29 Oct.-1 Nov Page(s): [8] Giuseppe Bianchi, Performance Analysis of the IEEE Distributed Coordination Function IEEE Journal on selected areas in communications,vol.18 No.3 March [9] B. P. Crow, I. Widjaja, J. G. Kim, and P. Sakai, Investigation of the IEEE Medium Access Control (MAC) Sub layer Functions. Department of Electrical and Computer Engineering,University of Arizona, IEEE 1997 Page(s): [2] C. Rama Krishna, Saswat Chakrabarti, and Debasish Dutta, A modified backoff algorithm for IEEE DCF based MAC protocol in a Mobile Ad-Hoc Network, TENCON IEEE Region 10 Conference, Volume B, Nov Page(s): Vol