File Latency Analysis

To compare the performance of MFTP, AMFTP, and NCFTP, a file of size 1MB is down-loaded from the server. Five transmission scenarios are considered in these experiments, including 1) Download one file via FTP by 3G only; 2) Download one file via FTP by WiFi only; 3) Download one file via MFTP by 3G and WiFi concurrently; 4) Download one file via AMFTP by 3G and WiFi concurrently; 5) Download one file via NCFTP by 3G and WiFi concurrently. We execute each scenario steps 100 times under limited WiFi bandwidth respectively 128KB/s, 192KB/s and 256KB/s. The file transfer latency of each transfer is recorded.

In Fig. 5.6, the average file transfer latency of our proposed protocols are shorter than FTP via WiFi only and 3G only. The average file transfer latency of MFTP, AMFTP,

and NCFTP under WiFi with 128KB/s are 11.654s, 11.355s, and 11.074s respectively. The average file transfer latency of MFTP are 57.07% and 52.36% of transfer latency of using WiFi only and 3G only. The ratio of AMFTP are 55.6% and 51.01% respectively. The ratio of NCFTP are only 54.23% and 49.76%. The average file transfer latency and bandwidth are given in Table 5.7.

WiFi 128 WiFi 192 WiFi 256

0

Figure 5.6: The average file transfer latency of five scenarios

We can see the performance of AMFTP better than MFTP, and the performance of NCFTP is better than AMFTP. The file transfer latency of NCFTP is 97.43% of AMFTP, and the transfer latency of AMFTP is 97.53% of MFTP. So NCFTP reduce the file transfer latency more than AMFTP. Similar results can be found in the experiments with WiFi bandwidth 192KB/s and 256KB/s. The results verify the performance of our proposed protocols. Our proposed protocols are successful shortened the file transfer latency. When the bandwidth of each interfaces are closed, the latency of our proposed protocols are nearly

Table 5.7: The average latency and bandwidth for a file of 1MB.

WiFi with bandwidth 128KB/s

3G WiFi MFTP AMFTP NCFTP

Time 22.25 20.42 11.65 11.36 11.07 KB/s 46.01 50.15 87.87 90.18 92.46

WiFi with bandwidth 192 KB/s

3G WiFi MFTP AMFTP NCFTP

Time 21.27 16.32 9.71 9.40 9.34 KB/s 48.15 62.74 105.38 108.88 109.69

WiFi with bandwidth 256KB/s

3G WiFi MFTP AMFTP NCFTP

Time 20.28 13.74 8.20 8.05 7.93 KB/s 50.51 74.51 124.89 127.26 129.17

the half latency of FTP via WiFi only or 3G only. We can see that there are no big problems for a mobile device to activate both radio interfaces. The bandwidth of MFTP and AMFTP are roughly equal to the sum of the bandwidth of the 3G and WiFi interfaces, but the bandwidth of NCFTP can further improve the performance. In Fig. 5.9, Fig. 5.8, and Fig.

5.7, the distribution of file transfer latency of FTP, MFTP, AMFTP and NCFTP are similar with our proposed math model in Fig. 3.1. It adequately verifies the truth of the math model. We find a phenomenon in Fig. 5.9 and Fig. 5.7. When the bandwidths of interfaces

are lower, the improvement of transfer latency of NCFTP is much bigger than AMFTP and MFTP.

5 10 15 20 25

0 20 40 60 80 100

second

%

NCFTP AMFTP MFTP FTP−WiFi FTP−3G

Figure 5.7: c.d.f of file transfer latency under WiFi with 256KB restriction

5 10 15 20 25

Figure 5.8: c.d.f of file transfer latency under WiFi with 192KB restriction

10 15 20 25 30

Figure 5.9: c.d.f of file transfer latency under WiFi with 128KB restriction

Chapter 6

Conclusions

In this work, we proposed three file transfer protocols for mobile devices equipped with multiple radio interfaces to reduce file transfer latency by integrating heterogeneous radio resources. We implemented the proposed protocols on a netbook to download files via WiFi and 3G connections, and observed performance improvement. In the experiment results, our proposed protocols are successful shortened the file transfer latency. The file transfer latency of our proposed protocols are almost 60% of transfer latency of FTP via single interface. When the bandwidth of each interfaces are closed, the latency of our proposed protocols are nearly half of transfer latency of FTP via single interface. When the bandwidth MFTP and AMFTP are roughly equal to the sum of the bandwidth of the 3G and WiFi interfaces, but the bandwidth of NCFTP can further improve the performance. The results verify the performance of our proposed protocols. Our proposed protcools are reducing the file transfer latency and raising the bandwidth utility of network interfaces. We also induce

the c.d.f math model of practical solution and ideal solution, and verify the truth of the math model. In the future, we hope the proposed protocols be used on file transfer, and we will develop general-purpose protocols not only for file transfer.

Bibliography

[1] [Online]. Available: http://sinwen.com/?p=4231

[2] [Online]. Available: http://www.eettaiwan.com/ART 8800580575 617723 NT d3c3a65f.HTM

[3] J. Zhang and R. D. McLeod, “A udp-based file transfer protocol with flow control using fuzzy logic approach,” in Electrical and Computer Engineering, 2003. IEEE CCECE 2003. Canadian Conference on, vol. 2, no. 2, 4-7 May 2003, pp. 827–830.

[4] K. Sklower, B. Lloyd, G. McGregor, D. Carr, and T. Coradetti, “The PPP multilink protocol (MP),” August 1996. [Online]. Available: http://www.rfc-editor.org/rfc/rfc1990.txt

[5] X. G. Wang, G. Min, J. Mellor, and K. Al-Begain, “A qos-based bandwidth management scheme in heterogeneous wireless networks,” International Journal of Simulations, 2004.

[6] Y. Choi and S. Choi, “Service charge and energy-aware vertical handoff in integrated ieee 802.16e/802.11 networks,” in INFOCOM 2007. 26th IEEE International Conference on Computer Communications., 6-12 May 2007.

[7] K. J. Peng and Z. Tsai, “Distortion and cost controlled video streaming in a heteroge-neous wireless network environment,” in The 17th Annual IEEE International Sympo-sium on Personal Indoor and Mobile Radio Communications, 2006, pp. 1–5.

[8] K. Chebrolu and R. Rao, “Bandwidth aggregation for real-time applications in hetero-geneous wireless network.pdf,” in Mobile Computing, IEEE Transactions on, 2006, pp.

388 – 403.

[9] K. Evensen, D. Kaspar, P. Engelstad, A. F. Hansen, C. Griwodz, and P. Halvorsen, “A network-layer proxy for bandwidth aggregation and reduction of ip packet reordering,”

in Local Computer Networks 2009. LCN 2009. IEEE 34th Conference on, 2009, pp. 585 – 592.

[10] J. C. Fernandez, T. Talebt, K. Hashimoto, Y. Nemoto, and N. Kato, “Multi-path scheduling algorithm for real-time video applications in next-generation wireless net-works,” in Innovations in Information Technology 2007. IIT 07. 4th International Con-ference on, 2007, pp. 73 – 77.

[11] D. Kaspar, K. Evensen, A. F. Hansen, P. Engelstady, P. Halvorsen, and C. Griwodz, “An analysis of the heterogeneity and ip packet reordering over multiple wireless networks,”

in Computers and Communications ISCC 2009 IEEE Symposium on, 2009, pp. 637 – 642.

[12] R. Ahlswede, N. Cai, S.-Y. R. Li, and R. W. Yeung, “Network information flow,” IEEE Transactions on Information Theory, vol. 46, no. 4, pp. 1204–1216, Jul 2000.

[13] S.-Y. R. Li, R. W. Yeung, and N. Cai, “Linear network coding,” IEEE Transactions on Information Theory, vol. 49, pp. 371–381, February 2003.

[14] S. Fong and R. Yeung, “Variable-rate linear network coding,” in Information Theory Workshop (ITW ’06), Oct. 2006, pp. 409–412.

[15] R. Koetter and M. M´edard, “An algebraic approach to network coding,” IEEE/ACM Transactions on Networking, vol. 11, no. 5, pp. 782–795, Oct. 2003.

[16] T. Ho, R. Koettert, M. M´edard, D. R. Karger, and M. Effros, “The benefits of coding over routing in a randomized setting,” in IEEE International Symposium on Information Theory (ISIT 2003), 29 June-4 July 2003, p. 442.

[17] S. Katti, H. Rahul, W. Hu, D. Katabi, M. M´edard, and J. Crowcroft, “Xors in the air:

Practical wireless network coding,” IEEE/ACM Transactions on Networking, vol. 16, no. 3, pp. 497–510, June 2008.

[18] H.-P. Wang, Y.-T. Chuang, C.-W. Yi, Y.-C. Tseng, and P.-C. Liu, “Xor-forwarding for wirelss networks,” in Global Telecommunications Conferences, (IEEE GLOBECOM 2009), 30 November - 4 December 2009.

[19] C. Gkantsidis and P. Rodriguez, “Network coding for large scale content distribution,”

in Proceedings of the 24th Annual Joint Conference of the IEEE Computer and Com-munications Societies (IEEE INFOCOM 2005), vol. 4, March 2005, pp. 2235–2245.

[20] [Online]. Available: http://en.wikipedia.org/wiki/Binomial distribution

[21] [Online]. Available: http://mathworld.wolfram.com/BinomialDistribution.html

[22] [Online]. Available: http://en.wikipedia.org/wiki/Negative binomial distribution

[23] [Online]. Available: http://mathworld.wolfram.com/NegativeBinomialDistribution.html

[24] E. Furman, “On the convolution of the negative binomial random variables,” Statistics

& Probability Letters, vol. 77, no. 2, p. 169V172, 15 January 2007.

[25] C.-L. Hwang and S.-Q. Li, “On input state space reduction and buffer noneffective region,” in Proceedings of the 13th Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE INFOCOM ’94), vol. 3, 12-16 June 1994.