Performance Analysis of CSMA/CA in Wireless Local Area Network

Transcription

1 International Journal of Coputer Applications ( ) Volue 120 No.10, June 2015 Perforance Analysis of CSMA/CA in Wireless Local Area Network Muhaad Younus Daffodil International University 102, Sukrabad, Mirpur Road Dhanondi, Dhaka Bangladesh A.K.M Falul Haque Daffodil International University 102, Sukrabad, Mirpur Road Dhanondi, Dhaka Bangladesh Md. Zahirul Isla Daffodil International University 102, Sukrabad, Mirpur Road Dhanondi, Dhaka Bangladesh ABSTRACT With the rapid advance of obile coputing, high speed Wireless Local Area Networks (LAN) attracts a lot of research interests in recent years. A new international stander IEEE 1 on wireless LAN has been established. Its physical transission is realied by either spread spectru counication or non-directive infrared. The ediu access control of IEEE 1 is using carrier sense ultiple access with collision avoidance (CSMA/CA) as the basic level access. Continuing fro the study of, we shall analye IEEE 1 CSMA/CA perforance with or without RTS/CTS in this paper. We have iproved the siulation by including RTS/CTS exchanges, which is part of the CSMA/CA protocol. Using this ethodology, we hoped to increase the nuber of successful transissions and decrease the nuber of collisions within the WLAN. This particular aspect of the siulation is deonstrated in the Analysis section of this work. Additionally, the siulation akes sure that once an RTS/CTS handshake is copleted; all other nodes not included in the data exchange reain silent. This is done to prevent nodes fro getting into a continuous cycle of RTS/CTS exchanges. General Ters Wireless Local Area Network, CSMA/CA Keywords IEEE 1, CTS/RTS 1. INTRODUCTION The world of today has becoe quite fast and reliable ainly because of the wireless counication. In our work we have studied on CSMA/CA and its perforance analysis in wireless local area network i.e. WLAN. We have siulated the code in MATLAB and get graphs by varying different paraeters. By providing enough reasons and explanation finally we would like to coent on the perforance of the CSMA/CA protocol and discuss the future aspect of it. With the siulation graphs and explanations, checking the respected theories we can conclude that using RTS/CTS in CSMA/CA protocol is efficient. Because in the WLAN we would like to get as uch higher the transission to get as possible and lower the collision aong the different nodes. We can see the clear evidence that we have given fro our siulation of the CSMA/CA protocol that without using RTS/CTS in the protocol gives the uch poor result in the transission and collision. But if we use RTS/CTS the transission increases as well as the collision decreases. And in the increased nuber of nodes the perforances of CSMA/CA protocol also increases for using the RTS/CTS condition. So the assuption of using RTS/CTS in the CSMA/CA protocol in every condition provides uch efficiency is proven here. 2. OVERVIEW OF WLAN All Wireless LAN is a eber of the IEEE 802 faily of specifications for Local Area Networks which allows coputers to get connected to a network through wireless[1]. And because of wireless is a shared ediu, everything that is transitted or received over a wireless network can be intercepted. Prot ocol Lega cy 1a 1b 1g 1n 1y Rele ase date Op. Frequ ency GH Throug hput (Typ) GH GH GH June June GH 5 GH 3.7 GH 4.3 Dat a Rat e (M ax) Modul ation Techni que Ran ge (Rad ius Indo or) ~20 OFDM ~35 DSSS ~38 OFDM ~38 ~70 ~50 Table 1: Characteristics of different version of wireless network protocols[1]. Rang e (Radi us Outd oor) ~100 ~120 ~140 ~140 ~250 ~5000 A survey had found that about 70% of wireless network have no security at all. Aong the IEEE 1b [2] is the ost popular and widely used protocol in Wireless area. So, we choose 1b and it s really iportant to ensure security in WLAN. There are two basic odes of operation specified in the standard. The ost coonly used ode is the infrastructure ode. The infrastructure ode allows for either one of the entities to be an access point. In ad-hoc ode all entities are considered clients. Ad-hoc ode ay also be referred to as independent ode. Stations in adhoc ode participate in an adhoc network, likewise if they are in infrastructure ode 43

2 International Journal of Coputer Applications ( ) Volue 120 No.10, June 2015 they participate in an infrastructure network. Interface of a client or access point contains a radio and an antenna. To avoid interference and allow networks to operate in the sae locations, IEEE 1 specifies grou of frequencies that ay be utilied by a network. Wireless LANs undoubtedly provides higher productivity and cost savings. In light of this, any organiations are beginning to deploy wireless LAN technologies not only for Cost savings, but also for convenience and flexibility of use. But the fundaental question plaguing the industry today is if wireless networks can be deployed securely without coproising organiation s assets - inforation. This study was undertaken to find out if wireless networks are inherently insecure thereby liiting enterprise deployent. If yes, what are the known holes, and can they be fixed? The following are the contributions to knowledge this work has ade through exhaustive and broad literature study[14]: 1. This study has shown that wireless LANs is prone to any different kinds of attacks. Attept to secure wireless LANs, suitable for enterprise deployent, initiated a ove fro weak WEP to ore robust WPA2. 2. This work also showed that the ost effective security solution for Wireless LANs involves a cobination of security technologies. 3. It deonstrated that a thorough vulnerability assessent and risk analysis is essential for developent of effective security policy and deterination of appropriate security easures, or cobination of easures that are ost effective. 4 It also showed that countereasures and best practices - like personal firewalls, antivirus, intrusion detection systes etc. - typically used to defend Internet connected Clients are generally recoended for WLAN clients as well. 5. On-going onitoring and periodic testing are necessary to verify that a deployed WLAN eets defined objectives. 3. CARRIER SENSE MULTIPLE ACCESS/COLLISION AVOIDANCE (CSMA/CA) In Wireless LANs CSMA/CA is core concept in counicating wirelessly. In any shared ediu accessing the ediu without collision is iportant part. Its like not talking all at a tie, so that reaining people should understood the other's talk. To know about CSMA/CA we have to know CSMA. Carrier Sense Multiple Access (CSMA) is one of ost popular rando access protocols in practice, which we see in ost of wireless textbooks. The key features of CSMA is that each link with a pair of transitter and receiver senses the ediu and transits a packet only if the ediu is sensed idle. Due to its siple and distributed nature, it has been regarded as one of the ost practical MAC protocols. in wireless networks, e.g., CSMA is a basic ediu access algorith in IEEE 1. Thus, there exists a vast array of research results on CSMA in ters of its analysis under various settings and its applications to practical systes. CSMA aintain a rule on a network that one point can transit data, otherwise it ay will loss of data/fraes. CSMA active at the tie when a point wants to send data or signal on a network. It always checks or listen for network traffic before transitting a signal or data frae, if it find any other transission is in process then it will pause its own process up to the previous process copleted. However, if ultiple devices access it siultaneously and a collision occurs, they both have to wait for a specific tie before reinitiating the transission process [7]. 3.1 Flow chart of CSMA/CA In CSMA/CA, when a point or node receives a packet that is to be sent, it checks for free tie period of a channel by which the node can sent the packet. In CSMA, a node or device have to wait for a tie interval if it find a busy or transferring process on the network. The tie period is known as back-off factor. Back-off factor is easure with back-off counter. When the transitting channel is clear, back-off counter reaches to ero and the device starts to send data/packet. And when the transitting channel is not clear but back-off counter is reaches to ero it eans that the process is repeated fro the sensing again the channel for rando tie.[10] [11] Fig 1: Flow chart of CSMA/CA[11] 3.2 Basic operation of CSMA/CA Now I discuss the basic operation of CSMA/CA device. This figure is the basic operation of CSMA/CA. Fig 2: Basic operation of CSMA/CA [9]. Before transitting, host A verifies that the channel is epty for a long enough period. Then, its sends its data frae[14]. After checking the validity of the received frae, the recipient sends an acknowledgeent frae after a short SIFS delay. Host C, which does not participate in the frae exchange, senses the channel to be busy at the beginning of the data frae. Host C can use this inforation to deterine how long the channel will be busy for. Note that as SIFS<DIFS<EIFS, even a device that would starts to sense the channel 44

3 International Journal of Coputer Applications ( ) Volue 120 No.10, June 2015 iediately after the last bit of the data frae could not decide to transit its own frae during the transission of the acknowledgeent frae. 3.3 Collision of CSMA/CA The ain difficulty with CSMA/CA is when two or ore device transits at the sae tie and cause collisions. This is illustrated in the figure below, assuing a fixed tieout after the transission of a data frae. With CSMA/CA, the tieout after the transission of a data frae is very sall, since it corresponds to the SIFS plus the tie required to transit the acknowledgeent frae. with either the sender or the reception of the data frae are infored of the reservation. They can copute the total duration of the transission and defer their access to the transission channel until then. This is illustrated in the figure below where host A reserves the transission channel to send a data frae to host B. Host C notices the reservation and defers its transission[4]. Fig 3: collision with CSMA/CA[6] To deal with this proble, CSMA/CA relies on a back off tier[6]. This back off tier is a rando delay that is chosen by each device in a range that depends on the nuber of retransissions for the current frae. The range grows exponentially with the retransissions as in CSMA/CD. The iniu range for the back off tier is where the slot Tie is a paraeter that depends on the underlying physical layer. Copared to CSMA/CD s exponential back off, there are two iportant differences to notice. First, the initial range for the back off tier is seven ties larger. This is because it is ipossible in CSMA/CA to detect collisions as they happen. With CSMA/CA, a collision ay affect the entire frae while with CSMA/CD it can only affect the beginning of the frae. Second, a CSMA/CA device ust regularly sense the transission channel during its back off tier. If the channel becoes busy (i.e.)because another device is transitting), then the back off tier ust be froen until the channel becoes free again. Once the channel becoes free, the back off tier is restarted. This is in contrast with CSMA/CD where the back off is recoputed after each collision. This is illustrated in the figure below. Host A chooses a saller back off than host C. When C senses the channel to be busy, it freees its back off tier and only restarts it once the channel is free again. We need soe syste configuration in our work.[10] [11] 3.4 How to avoid collision in CSMA/CA Do For avoiding collision in CSMA/CA, every device can reserve the operating channel for short period of tie. The exchange of two control fraes- Request to Send(RTS) and Clear to Send(CTS) are used to avoid the risk of collisions between different node in the network. For reserving a channel, first one device send RTS frae to intended destination device. this RTS frae contains the inforation about reservation period. The other devices replies by a CTS control frae which also contain inforation about duration of channel reservation. As the duration of the reservation has been sent in both RTS and CTS, all hosts that could collide Fig 4: Avoid collision in CSMA/CA[4]. not include headers, footers or page nubers in your subission. These will be added when the publications are assebled. 4. SIMULATION AND RESULTS To analye the CSMA/CA protocol we have created a siilar siulation environent considering the transission of different nodes. And then we have siulated the protocol, in this siulation we have varied different variables. These variables are 1. CSMA/CA protocol with RTS/CTS 2. CSMA/CA protocol without RTS/CTS 3. Changing the no. of nodes for both cases. So far what we have got fro the siulation are included in this point. We would like to explain the results with graphical representation. By providing enough reasons and explanation finally we would like to coent on the perforance of the CSMA/CA protocol and discuss the future aspect of it. Successful transission when the no. of node is less or equal 5 First of all here is the graph of successful transission when the no. of node is less or equal 5. Figure 5: Successful transissions with and without rtscts when node nuber is 5 45

4 International Journal of Coputer Applications ( ) Volue 120 No.10, June 2015 Siulation Data Transission with rts/cts Transission without rts/cts Table 2: Siulation Data (transissions) for when node nuber is 5 Here the X-axis denotes the nuber of nodes and Y-axis stands for average nuber of data packets. Fro the graph we can see two different color of lines. The red line represents the transission with RTS/CTS and blue one is for transission without RTS/CTS. This graphical representation is enough to coent that if we use RTS/CTS in CSMA/CA protocol the successful transission is higher than the other one. Average nuber of data packets value is given in a table also, fro which we can conclude the sae. Bar-Chart: For a better understanding we can use the bar chart and here it is Collision with rts/cts Collision without rts/cts Siulation Data Table 3: Siulation Data (collision) for when node nuber is 5 In the graph the axis's denote the sae again and blue or red color indicates the sae as before. If we use RTS/CTS then the collision decreases that we can see fro the graph. For the nuber of node respected values are given also in the tabular for. If we want to see the collision difference for these two protocols with respect to individual node point this bar chart provides uch ore clear explanation Bar-Chart Figure 6: Successful transissions with and without rtscts when node nuber is 5(Bar-Chart) The green color represents the transission with RTS/CTS and the red is for the other protocol. Fro here we can clearly see that the difference between this two protocol with respect to average nuber of data packets is uch higher. And the transission with RTS/CTS wins over the other. Collision rate when the no. of node is less or equal 5 Though we can conclude that using RTS/CTS in CSMA/CA is better by seeing the transission graph, we would like to see what does happen in the collision Figure 8: Collision rate with and without rts-cts when node nuber is 5(Bar-Chart) For every single node the red one that eans collision without RTS/CTS in the CSMA/CA protocol gives higher values. So in the sense of collision using RTS/CTS in the protocol provides better result since in here collision is less, so transission would be higher. Successful transission when increase the nuber of nodes Now we would like to increase the nuber of nodes and see what happens in the transission and collision graph. First the successful transission graph Figure 9: Successful transissions with and without rtscts in when node nuber is increasing. Siulation Data Figure 7: Collision rate with and without rts-cts when node nuber is 5 46

5 International Journal of Coputer Applications ( ) Volue 120 No.10, June 2015 Transission with rts/cts Transission without rts/cts Table 4: Siulation Data (transission) for when node nuber is increasing Here the nuber of node is increased fro 5 to 10. The graphical paraeters are sae as before. Since fro the graph we can understand that the red line has higher values in the transission process we can conclude that the RTS/CTS is better than the another protocol. With the increased nuber of nodes the transission continuously decreases for without RTS/CTS but using RTS/CTS can get better packet transission. The tabular data for successful transission supports the sae conclusion. Bar-Chart For a clear understanding in each of the nodes this bar chart hel to shade a brighter light Siulation Data Collision with rts/cts Collision without rts/cts Table 5: Siulation Data (collision) for when node nuber is increasing. Here we can clearly see that the blue line eans the collision without RTS/CTS is higher than the red line eans collision with RTS/CTS. If the collision is lower the better so the blue line refers that not using RTS/CTS in the CSMA/CA protocol get the worse result. Bar-Chart Again if we would like to get the clear picture what really happens in the each node this bar chart will help to understand Figure 10: Successful transissions with and without rtscts in when node nuber is increasing(bar-chart) Here the green one is for RTS/CTS which is higher in every node than the without RTS/CTS. The higher the better transission, so RTS/CTS has better perforance on the increased nuber of nodes. Collision rate when increase the nuber of nodes: If we do not use RTS/CTS in the CSMA/CA protocol for the increased nuber of nodes the collision would increase for each increased nuber of node. This graph of collision provides the sae result Figure 12: Collision rate with and without rts-cts in Bar- Chart when node nuber is increasing. With the increased nuber of nodes in each node the collision without RTS/CTS is uch higher than the collision with RTS/CTS. We can su up that the higher the collision the worsen the perforance, so without using RTS/CTS will provide the poor perforance in the CSMA/CA protocol. 5. ANALYSIS With the siulation graphs and explanations, checking the respected theories we can conclude that using RTS/CTS in CSMA/CA protocol is efficient. Because in the WLAN we would like to get as uch higher the transission to get as possible and lower the collision aong the different nodes. We can see the clear evidence that we have given fro our siulation of the CSMA/CA protocol that without using RTS/CTS in the protocol gives the uch poor result in the transission and collision. But if we use RTS/CTS the transission increases as well as the collision decreases. And in the increased nuber of nodes the perforances of CSMA/CA protocol also increases for using the RTS/CTS condition. So the assuption of using RTS/CTS in the CSMA/CA protocol in every condition provides uch efficiency is proven here. Figure 11: Collision rate with and without rts-cts in when node nuber is increasing 47

6 International Journal of Coputer Applications ( ) Volue 120 No.10, June CONCLUSION With the siulation graphs and explanations, checking the respected theories we can conclude that using RTS/CTS in CSMA/CA protocol is efficient. Because in the WLAN we would like to get as uch higher the transission to get as possible and lower the collision aong the different nodes. We can see the clear evidence that we have given fro our siulation of the CSMA/CA protocol that without using RTS/CTS in the protocol gives the uch poor result in the transission and collision. But if we use RTS/CTS the transission increases as well as the collision decreases. And in the increased nuber of nodes the perforances of CSMA/CA protocol also increases for using the RTS/CTS condition. Thus the assuption of using RTS/CTS in the CSMA/CA protocol in every condition provides uch efficiency is proven in this part. 7. FUTURE WORK For this way when RTS and CTS are used, then huge nuber of frae or data is saved and collision can be avoided. So, we will try to better solution in future that will decrease data or packet loss and achieve highly otivated the carrier sense ultiple access with collision avoidance. 8. REFERENCES [1] ANSI/IEEE Standard 1. Piscataway, NJ: LAN MAN society, IEEE Coputer Society.Copyright Available online [2] Li, Keyong and Sherai, Mohaed. Wireless MAC 1. Unpublished work, Final paper for SC 546, Fall Available online at sedocuents/a349b830b c631d6136de de87/paper/final paper.doc. [3] IEEE 1 WG, Part 11: Wireless LAN Mediu Access Control (MAC) and Physical Layer (PHY) Specifications, IEEE Std. 1, 1999 [4] IEEE 1 WG, Part 11: wireless LAN ediu access control (MAC) and physical layer (PHY) specifications: ediu access control (MAC) quality of service (QoS) enhanceents, IEEE 1e D8.0, Feb [5] Carrier Sense Multiple Access with Collision Avoidance, Available online at [6] Wireless LAN Networking White Paper. (for CSMA/CA) [7] Working principle of CSMA/CA. csaca-works.htl (working principle of CSMA/CA) [8] Throughput Modelling and Fairness Issues In CSMA/CA Based Ad-Hoc Networks. By Xin Wang and Koushik Kar. rs/infoco05-csa.pdf [9] CSMA/CA htt:// utf-8&oe=utf-8&rls=org.oilla:en- US:official&client=firefoxa&gws_rd=cr&ei=d8OtU9WAJ9WMuATltoKoCw CSMA/CA [10] C. H. Yeh, A collision-controlled MAC protocol for obile adhoc networks and ultihop wireless LANs to be appeared in Proceeding of IEEE Global Telecounications Conference, [11] Source URL: htt://sc.info.ucl.ac.be/release/cnp3/book/0.2/htl/lan /lan.htl#carrier-sense-ultiple-access-with-collision Avoidance. [12] CS402: COMPUTER COMMUNICATIONS AND NETWORKShttp:// uter Science 402) [13] Wireless Mediu Access Control (MAC) and physical layer (PHY)specifications: Aendent: ESS Mesh networking. IEEE P1s/D1.00(Nov. 2006). [14] Sunday, N. (2008). Wireless Local Area Network (WLAN): Security Risk Assessent and Countereasures. Blekinge Institute of Technology School of Engineering, (August). [15] C. H. Yeh, A collision-controlled MAC protocol for obile adhoc networks and ultihop wireless LANs to be appeared in Proceeding of IEEE Global Telecounications Conference, [16] S. Ray, J. B. Carruthers, and D. Starobinski, RTS/CTSinduced congestion in ad hoc wireless LANs, WCNC, IJCA TM : 48