Upload Traffic over TCP and UDP Protocols in Different Security Algorithms in Wireless Network

Upload Traffic over TCP and UDP Protocols in Different Security Algorithms in Wireless Network Abdalla Gheryani, And Mladen Veinović Abstract This paper studies and measures the outcome of different security algorithms on the performance of a wireless LAN over TCP and UDP protocols. Real experiments were performed on a wireless LAN by uploading traffic in different data rate. The data obtained was analyzed for throughput, jitter and delay under different security scenarios. Both TCP and UDP traffic streams were analyzed at three different data rates. The effect of congestion is also measured. The results tell that no important degradation in performance occurs by enabling security algorithms in a wireless LAN. In low bit rate, it shows there is no important degradation, but in case of high rate more than channel rate, it shows there is influence. Keywords Wireless LAN, TCP, UDP, WPA, WEP. W I. INTRODUCTION IRELESS networks have extended excellent acceptance in marketable, military, educational and research in last few years. Mobility support is another salient feature of wireless networks which grants the users not only anytime, anywhere network access, but also the freedom of roaming while networking. The main contributors to this acceptance are flexibility and mobility offered by these networks. The significant dependence on wireless networks in all walks of life has created a tremendous need for increasing the reliability and security of these networks. The security risks in wireless networks are more than those of wired networks due to open access of the shared wireless medium [2]. Besides security, performance is another major issue in wireless networks. These issues have been exclusively studied in an extensive manner. However, little work exists in the area of understanding the relationship between these two issues. The aim of our study is to understand and quantify the relationship between security and performance in wireless LANs (Local Area Networks). To carry out this study, we performed experiments on a wireless LAN by enabling security protocols like WEP and WPA for TCP and UDP traffic at data rates of 1 Mbps, 5Mbps and 14 Mbps. To see the impact of these Abdalla Gheryani is with the Singidunum University, Belgrade, CO 11000 Serbia (e-mail: abdalla.gheryani@singimail.rs). Mladen Veinović, is now Vice Rector of University, Singidunum University, Belgrade, CO 11000 Serbia (e-mail: mveinovic@ singidunum.ac.rs). security protocols, experiments were performed in unencrypted case as well. The data obtained from these experiments is compared for performance metrics like throughput, jitter and delay. II. RELATED WORK Some work has been done in this direction and has revealed differing results [2]-[6]. In [2], authors have examined the effect of WEP on throughput in adhoc networks. The results indicate a decrease in throughput in the presence of security. In [3], the authors compared the strength of various security protocols and the overheads involved in implementing them. The results indicate an increase in overhead with the increase in security strength. The throughput and delay are affected differently by different security policies. In [5], the author has studied performance of wireless LANs for multiple clients. The results indicate degradation in performance with the increase in number of clients and also with the increase in security strength. In [6], the author investigated the performance and security issues of IEEE 802.11 wireless networks using the IEEE 802.1X and Virtual Private Network (VPN) models. The results obtained showed that the VPN model impacted the performance more than the 802.1X model. The survey reveals that most of the researchers have studied the impact of implementing security on throughput. Therefore, in this paper, an attempt has been made to study the impact of security on these performance parameters in congested and uncongested networks. III. NETWORK LAYOUT AND PROCEDURE In our scenario, we have used wired node N0 (Dell with Pentium Dual-Core 2.2 GHz, Marvell Yukon Fast Ethernet Controller and Windows 7 Ultimate with service pack1) as the receiver and wireless node N1 (Compaq with Pentium Core Duo 1.83 GHz, Broadcom Wireless LAN and Windows 7 Ultimate with service pack1) as the sender. In the topology R1 (Cisco 2100 Cable Modem) act as Cisco Modem while R2 (Linksys E1500 Wireless-N Router with SpeedBoost) is Cisco Access Point. Ethernet node N0 connects with R1 through a 100 Mbps link. The link bandwidth between R1 and R2 is set to 100 Mbps while the wireless link between R2 and N1 operates at a nominal data rate 11 Mbps. 320

TABLE I BRIEF DESCRIPTION OF DIFFERENT SECURITY SCENARIOS Security Label Explanation Scenario No Security S1 No Encryption WEP-64 S2 WEP Protocol with 64 bit key WEP-128 S3 WEP Protocol with 128 bit key WPA-TKIP S4 TKIP is used for data encryption Fig. 1 Layout of the Network The experiments were conducted on wireless test for different security scenarios and traffic streams in the infrastructure mode of wireless LANs. A brief discussion of different aspects of these experiments is given below. IV. SECURITY SCENARIOS The experiments were carried out for following scenarios:- No Security: In this scenario the entities communicate over wireless link without any authentication and encryption. The results obtained are used as a reference for comparison with security enabled cases. WEP enabled: In this scenario WEP encryption is enabled. The experiments were performed for 64 bit and 128 bit key sizes. WPA enabled: In this scenario experiments were performed using TKIP and AES mode supported by the access point. V. TRAFFIC STREAM TCP and UDP traffic streams were chosen for experiments. The traffic was generated and received using pathtest tool, installed on both communicating objects. VI. BANDWIDTH The Access Point used in the experiments can support data rates up to 54 Mbps. The data rate of Access Point (transmission rate of the wireless channel) was fixed at 11 Mbps. The data generation rate for the source was chosen at different transmission rate. These values are labeled as outgoing bandwidth. Since the data rate of Access Point was fixed at 11 Mbps so the generation rate less than 11 Mbps simulates the behavior of an uncongested network, whereas more than 11 Mbps generation rate represents a congested network. Table 1, shows brief description different security scenarios: WPA-AES S5 AES is used for data encryption VII. DATA COLLECTION For data collection following points were considered. To allow the test to be stabilize, first two readings were unwanted. Each experiment was carried out 5 times for reliability of data. The readings were noted down when we were sure that. For each experiment mean of the readings taken was computed. VIII. EXPERIMENT RESULT We present the results obtained from experiments. The experimental data for average throughput of TCP and UDP, delay of TCP and UDP and jitter of TCP and UDP are tabulated and compared for different scenarios. IX. AVERAGE THROUGHPUT For comparison purposes the data is plotted using OriginPro as shown in Figures 2, 3, 4, 5, 6 and 7. Following observations are made:- There is no major degradation in average throughput by enabling security policies like WEP-64, WEP-128, WPA- TKIP and WPA-AES. For 1and 5 Mbps traffic, the average throughput is near to the source data rates. However, for 12 Mbps case, the average throughput is less than the source data rate. This is due to congestion as the bandwidth of wireless channel is fixed at 12 Mbps. Fig. 2 Average Throughput of TCP in 1 Mbps 321

Fig. 3 Average Throughput of UDP in 1 Mbps Fig. 4 Average Throughput of TCP in 5 Mbps Fig. 5 Average Throughput of UDP in 5 Mbps Fig. 6 Average Throughput of TCP in 14 Mbps Fig. 7 Average Throughput of UDP in 14 Mbps The above figures show the achieved throughput for the two transport protocols, TCP and UDP as a function of the transmission rate. It can be seen that TCP and UDP achieve almost the same throughput for a send rate of 1 Mbps to 5 Mbps. For transmission rates, beyond 12 Mbps to 14 Mbps, TCP achieved 4.3 Mbps to 8.6 Mbps, whereas UDP achieved 4.1376 Mbps to 7.9008Mbps. Experiments show that there is no significant decrease in average throughput for various security scenarios. For 1 and 5 Mbps cases, no effect of congestion was seen as the average throughput is almost same as the data transmission rate at the source. The results obtained for 14 Mbps case show that the average throughput is less than the transmission rate. This is attributed to the congestion caused in the network by high generation rate. X. JITTER The jitter performance for each of the two transport protocols is shown in Table II, III and IV. The jitter performance for both protocols is depicted in the above tables. It is observed that for TCP, jitter values range from 0.057 ms to 76.261 ms as the transmission rate varies from 1 to 14 Mbps with difference security scenario. For UDP, the jitter values lie in the range from 0.057 ms to 12.85 ms. It can be noticed that with AES algorithm exhibit better performance in comparison with other security scenario. As transmission rate increases, for TCP, we can noticed that the jitter value has taken different values less than the values of low transmission rate depends on the security algorithm that has been used and for UDP, we can notice that the jitter values take lower values with increasing in the transmission rate. XI. DELAY Delay refers to the time taken for a packet to be transmitted across a network from source to destination. Tables V, VI and VII show the delay performance for each of TCP and UDP protocols. As the transmission rate is increased from 1 Mbps to 14 Mbps, delay experienced with TCP as transport protocol varies from 0.057 ms to 194.411 ms, while for UDP the delay is in the range of 0.2 ms to 94.548 ms. We can observe that, when using the TCP protocol the delay is taken varies difference value even with using different security algorithm. In UDP, the delay is slightly same and the starting value of delay is increased with changing of transmission rate and at transmission rate of 14 Mbps, the 322

delay is taken the highest value of 94.548 ms, because of congestion. XII. CONCLUSIONS We quantified the effect of implementing security on network performance in infrastructure mode wireless LAN. Experiments were performed under different security scenarios for TCP and UDP traffic streams. The security scenarios chosen were Wired Equivalent Privacy (WEP) with key sizes of 64 and 128 bits and Wi-Fi Protected Access (WPA) using TKIP and AES algorithms. For comparison purposes experiments were also performed for unencrypted case as well. The data rate for AP (potential bandwidth of wireless channel) was selected as 11 Mbps. The results obtained from experiments show that there is no different in average throughput for unencrypted and encrypted scenarios. UDP traffic shows there is slightly difference between unencrypted and encrypted value, because of congestion. For jitter performance, TCP is taken different values and that happen with changing of the encrypted algorithm and in UDP the jitter value is increasing in all encrypted scenario. For delay performance, there are varies values the delay had taken with TCP protocol in both unencrypted and encrypted algorithms, but when using UDP protocol, the delay almost same with all scenario. REFERENCES [1] Abdalla Gheryani, Mladen Veinović, Quantification of the Different Security Algorithms in Wireless Network, IJCSNS International Journal of Computer Science and Network Security, Vol. 12 No. 6 pp. 18-27, June 2012. [2] H. Yang, F. Ricciato, S. Lu, and L. Zhang, Securing a wireless world, Proc. IEEE (Special Issue on Cryptography and Security Issues), vol. 94, no. 2, February 2006. [3] A. K. Agarwal and W. Wang, Measuring performance impact of security protocols in Wireless Local Area Networks, The Second International Conference on Broadband Networks, Boston, USA, [4] October 2005. B. Smith, An approach to graphs of linear forms (Unpublished work style), unpublished. [5] M. Boulmalf, E. Barka, and A. Lakas, Analysis of the effect of security on data and voice traffic in WLAN, Computer Communications, 30 (2007) 2468-2477. [6] N. Baghaei, IEEE 802.11 wireless LAN security performance using multiple clients, University of Canterbury, Christchurch, NZ. C. J. Kaufman, Rocky Mountain Research Lab., Boulder, CO, private communication, May 1995. [7] J. Wong, Performance investigation of secure 802.11 wireless LANs: Raising the security bar to which level? University of Canterbury, Christchurch, NZ.M. Young, the Technical Writers Handbook. Mill Valley, CA: University Science, 1989. [8] A. M. Al Naamany, A. Al Shidhani, and H. Bourdoucen, IEEE 802.11 wireless LAN security overview, International Journal of Computer Science and Network Security, vol.6, no. 5B, May 2006. [9] W. Arbaugh, N. Shankar, Y. Wan, and K. Zhang, Your 802.11 wireless network has no clothes, IEEE Wireless Communication, vol. 9, no. 6, Dec. 2002. [10] W. Stallings, Cryptography and Network Security: Principles and Practice, 4th ed., Published by Dorling Kindersley (India) Pvt. Ltd., Licensees of Pearson Education in South Asia. Abdalla Gheryani was born on February 27, 1973. Received the B.S. in Computer Science and Engineering from Engineering Academy, Tajura - Libya in 1999 and M.S. degrees in Computer Science and Engineering from Jaypee Institute of Information Technology, Noida - India in 2009. Now pursing PhD. at Singidunum University, Belgrade Serbia. Mladen Veinović was born on January 01, 1962. He received the B.Sc., M.Sc. and Ph.D. degree in 1986, 1990 and 1996, respectively, all from the Faculty of Electrical Engineering, University of Belgrade. Since 1987, he has worked at the Institute of Applied Mathematics and Electronics. Since 2005, he is professor at Singidunum University. His current research interests include computer network, databases and data security. TABLE II JITTER FOR 1 MBPS 1 53.06 4.658 41.06 3.857 51.66 5.657 53.06 5.857 76.261 4.657 2 6.0576 3.257 47.06 0.258 26.059 0.858 46.66 0.458 11.058 6.058 3 5.8554 0.257 5.253 1.657 4.054 0.657 11.853 0.657 7.654 3.857 4 42.6594 2.457 43.059 0.657 45.46 1.457 49.66 0.657 52.06 0.457 5 6.7786 1.057 11.456 0.257 3.854 0.857 14.653 0.657 5.253 1.857 6 14.5356 3.058 9.257 1.057 13.658 0.657 32.859 1.257 38.259 0.857 7 23.6586 0.257 44.46 1.258 48.06 0.858 27.659 1.257 48.26 0.457 8 14.6528 0.857 9.254 1.657 12.255 0.257 10.454 0.658 10.053 0.858 9 43.6594 0.257 19.458 1.057 16.658 0.257 45.26 0.857 47.46 1.857 10 63.8826 1.755 41.283 0.945 16.482 0.515 37.882 0.825 63.282 0.975 323

TABLE III JITTER FOR 5 MBPS 1 0.057 3.257 0.457 3.258 14.858 5.057 2.458 2.057 8.458 7.858 2 3.658 1.658 1.057 0.257 0.457 2.658 0.657 0.658 0.856 0.256 3 5.657 0.657 5.058 1.057 1.457 0.657 1.458 1.857 1.657 0.658 4 0.658 1.057 1.257 0.257 0.857 1.257 1.056 1.057 0.858 3.057 5 1.556 1.057 2.057 1.257 0.457 0.457 0.857 1.257 1.057 0.857 6 0.957 1.057 0.257 2.458 2.458 1.657 0.857 6.858 1.057 0.657 7 5.658 0.458 1.257 1.056 1.257 0.257 1.858 2.857 1.857 3.058 8 1.257 1.057 2.058 0.458 0.657 0.658 1.457 0.857 1.258 1.057 9 3.457 0.857 0.057 1.857 2.257 0.457 2.057 0.457 1.457 0.857 10 1.085 1.475 0.685 1.035 0.885 1.875 1.685 0.765 0.885 0.885 TABLE IV JITTER FOR 14 MBPS 1 1.257 12.858 0.857 1.057 4.458 2.461 0.657 6.858 0.457 1.057 2 0.458 3.057 0.459 1.257 2.456 1.453 0.859 0.457 0.857 3.457 3 0.257 2.057 5.857 1.658 1.657 6.258 1.456 0.657 1.657 0.46 4 0.459 4.857 1.656 4.657 0.459 3.059 0.457 0.657 0.657 0.256 5 0.655 2.057 1.658 2.258 1.456 1.857 2.457 6.857 0.859 0.856 6 6.658 1.657 1.056 2.062 0.657 3.46 0.26 0.458 0.656 0.457 7 0.656 1.257 1.058 1.054 0.858 1.653 1.655 0.057 0.857 0.657 8 0.659 1.858 0.856 2.659 1.857 2.456 0.858 2.457 2.858 2.058 9 1.456 1.057 2.058 1.454 0.259 1.458 1.257 0.857 1.057 1.057 10 0.885 1.685 1.285 2.684 0.883 2.085 0.684 4.285 0.085 1.285 TABLE V DELAY FOR 1 MBPS 1 53.06 4.658 41.06 3.857 51.66 5.657 53.06 5.857 76.261 4.657 2 59.1176 7.915 88.12 4.115 77.719 5.515 99.72 5.315 87.319 10.715 3 24.973 6.172 13.373 3.772 21.773 6.172 21.573 4.972 24.973 14.572 4 67.6324 8.629 56.432 4.429 67.233 5.629 71.233 4.629 77.033 5.029 5 194.411 7.686 37.888 4.686 21.087 5.486 5.886 5.286 2.286 4.886 6 58.9466 10.744 27.145 3.743 34.745 5.143 38.745 4.543 40.545 5.743 7 82.6052 8.001 71.605 5.001 82.805 5.001 66.404 4.8 88.805 5.2 8 7.258 8.858 20.859 3.658 35.06 5.258 6.858 4.458 8.858 5.058 9 50.9174 7.115 40.317 4.715 51.718 5.515 52.118 4.315 56.318 4.915 10 114.8 4.87 81.6 3.66 68.2 2.03 90 2.14 119.6 1.89 324

TABLE VI DELAY FOR 5 MBPS 1 0.057 3.257 0.457 3.258 14.858 5.057 2.458 2.057 8.458 7.858 2 3.715 2.915 1.514 3.515 0.315 7.715 1.115 2.715 3.314 1.114 3 9.372 3.572 6.572 1.572 1.772 4.372 19.573 4.572 1.971 1.772 4 10.03 4.629 2.829 1.829 12.629 1.629 2.629 2.629 2.829 4.829 5 3.586 2.686 4.886 3.086 2.086 1.086 2.486 1.886 0.886 3.686 6 0.543 3.743 3.143 5.544 4.544 0.743 1.343 10.744 1.943 2.343 7 6.201 4.201 1.4 1.6 2.801 1 3.201 13.601 3.8 5.401 8 2.458 5.258 3.458 2.058 1.458 1.658 1.658 4.458 5.058 3.458 9 5.915 4.115 3.515 3.915 3.715 2.115 3.715 4.915 4.515 2.315 10 7 1.59 1.2 2.95 0.6 1.99 5.4 3.68 1.4 0.2 TABLE VII DELAY FOR 14 MBPS 1 1.257 12.858 0.857 1.057 8.458 67.461 0.657 6.858 0.457 1.057 2 1.715 15.915 24.316 2.314 0.914 0.914 25.516 3.315 1.314 4.514 3 1.972 11.972 34.173 0.972 0.571 7.172 1.972 3.972 0.971 38.974 4 33.431 16.829 1.829 5.629 25.03 47.231 0.429 0.629 1.628 20.23 5 2.086 1.886 20.487 17.887 1.486 41.088 2.886 7.486 26.487 2.086 6 8.744 0.543 0.543 4.949 1.143 94.548 42.146 7.944 3.143 2.543 7 2.4 1.8 13.601 40.003 19.001 4.201 0.801 3.001 1 2.2 8 23.059 1.658 1.457 74.662 1.858 2.657 19.659 5.458 21.858 7.258 9 1.515 0.715 3.515 27.116 43.117 14.115 16.916 1.315 5.915 1.315 10 0.4 2.4 16.8 8.8 1 2.2 0.6 5.6 2 0.6 325