Embedded Files
Abstract
As computer vision projects continue to spread with the popularization of easy-to-train neural networks such as YOLOv4 (you only look once), video can be captured with entry-level experience and uploaded to a server with GPUs (graphics processing units) where video processing can occur. However, these video files should not be streamed with the traditional unreliable data transfer of UDP (user-defined protocol) without reliability and congestion control added in the application layer as losses are not tolerable in preprocessed data. This project seeks to test the trade-off between goodput/transfer time and physical client mobility when transferring large files simultaneously from multiple cameras at varying distances utilizing TCP (transfer control protocol). Specifically, this project is targeted at Eye-bex camera modules (a current capstone project at SFU), which will be placed inside a climbing gym to acquire and upload video footage of rock climbers to a server for video processing.
To simulate the experiments, Riverbed Modeler will be used with a network topology of 2 cameras performing FTP applications of the average video chunk size, 53MB, every 15 minutes. Five wireless stations representing the expected traffic of light browsing from users are included in the network topology to account for simultaneous traffic.
References to Related Works
[1] D. Duchamp and N. Reynolds, "Measured Performance of a Wireless LAN", LCN, pp. 494-499, 1992. [Accessed 15 October 2021].
[2] N. Islam, C. C. Bawn, J. Hasan, A. I. Swapna, and M. S. Rahman, “Quality of service analysis of ethernet network based on packet size,” Journal of Computer and Communications, 18-Mar-2016. [Online]. Available: https://www.scirp.org/journal/paperinformation.aspx?paperid=65356. [Accessed: 15-Oct-2021].
[3] Y. Peng, H. Wu, K. Long and S. Cheng, "Simulation analysis of TCP performance on IEEE 802.11 wireless LAN," 2001 International Conferences on Info-Tech and Info-Net. Proceedings (Cat. No.01EX479), 2001, pp. 520-525 vol.2, doi: 10.1109/ICII.2001.983631.
[4] Y. Shui “Networking for big data,” Google Books, pp. 139-158. [Online]. Available: https://books.google.ca/books?hl=en&lr=&id=X2BECgAAQBAJ&oi=fnd&pg=PA139&dq=ftp%2Blan%2Bethernet%2Bpacket%2Bloss%2Bspeed&ots=tlHrBMCqhd&sig=5_ZGSxGaMOCbdqsv6rlRG6OvK24#v=onepage&q&f=false. [Accessed: 15-Oct-2021].
[5] J. A. R. P. de Carvalho, H. Veiga, P. A. J. Gomes, C. F. R. Pacheco, N. Marques, and A. D. Reis, ‘WI-FI Point-to-Point Links: Performance Aspects of IEEE 802.11a, b, g Laboratory Links’, in Electronic Engineering and Computing Technology, S.-I. Ao and L. Gelman, Eds. Dordrecht: Springer Netherlands, 2010, pp. 507–514. doi: 10.1007/978-90-481-8776-8_43.
[6] P. Das, B. Purkayastha and A. Debnath, "Large Size File Transfer over Wireless Environment", 2010 Second International Conference on Computer Engineering and Applications, 2010. Available: 10.1109/iccea.2010.281 [Accessed 15 October 2021].
Page updated
Google Sites
Report abuse