Login

An investigation of tools and techniques to measure/identify end-to-end bandwidth bottlenecks in the Internet

VA
State: completed by Jan Meier
Published: 2011-11-02

Still today, network bandwidth is a critical resource in the Internet because of heterogeneous bandwidth channels, access technologies, file sizes, and different requirements of different applications. Therefore, this heterogeneity and different requirements, for example, can lead unaware application to stream a 5GB video file over a 19.2Kb/s cellular data link, or send a text-only version of a web site over a 100Mb/s dedicated link. Or, even worse, an application delivering critical content to its customers in a region where there are recurrent bandwidth bottlenecks or performance degradation related to the Internet links along the routing path. Therefore, knowledge of the bandwidth along a path allows an application to avoid such mistakes by adapting the behavior and quality of its content, by choosing the best way to access/deliver some content, or yet, by choosing the most beneficial geographical places for some mirrors.

There are many tools like, pchar [1], nettimer [2], pathneck [3, 4, 5], YaZ [6], pathchar [7], and others [8, 9, 10], that measure and/or identify bandwidth bottlenecks along the path of two nodes in the Internet. Each of these tools use different techniques to measure and identify it, even though all of them use a traceroute-based method to identify the location of which is the "problematic" router in terms of bandwidth. Some of these techniques prioritize accuracy but having a long time of measurement (also consuming a lot of bandwidth), and others prioritize a faster measurement (less time, less bandwidth) but tending to be less accurate (although, trying to reach an acceptable level of accuracy). Moreover, some techniques are more intrusive than others, since they assume to have control over the two nodes (i.e., the end nodes), and others assume to be less intrusive but performing the measurement just from one of the end nodes (one-way).

Therefore, this thesis aims to study all possible tools and its techniques found in the literature. The student should install the tools, execute them, perform some measurements, and compare the results. The student should compare the techniques to justify the practical results (and its differences), pointing the cons and pros of each tool/technique. If techniques are found without a proof-of-concept tool, the student should do a brief comparison based on the evaluation presented in the scientific papers. Depending on available time and agreed work type, the student can have the chance to propose enhancements for such techniques or tools, implementing it, and evaluating such proposal.

References:

[1] Pchar tool website. Available at: http://www.kitchenlab.org/www/bmah/Software/pchar/. Last visited on: November, 2011.

[2] Nettimer: A Tool for Measuring Bottleneck Link Bandwidth. Available at: http://nms.lcs.mit.edu/~dina/MNM/mmdocs/nettimer.pdf. Last visited on: November, 2011.

[3] Pathneck tool website. Available at: http://www.cs.cmu.edu/~hnn/pathneck/. Last visited on: November, 2011.

[4] N. Hu, L. E. Li, Z. M. Mao, P. Steenkiste, and J. Wang: Locating Internet Bottlenecks: Algorithms, Measurements, and Implications. In SIGCOMM, 2004. Available at: http://www.cs.cmu.edu/%7Ehnn/papers/bottleneck.pdf. Last visited on: November, 2011.

[5] N. Hu, L. E. Li, Z. M. Mao, P. Steenkiste, and J. Wang: A Measurement Study of Internet Bottlenecks. In INFOCOM, 2005. Available at: http://www.cs.cmu.edu/%7Ehnn/papers/infocom05.pdf. Last visited on: November, 2011.

[6] J. Sommers, P. Barford, and W. Willinger: A Proposed Framework for Calibration of Available Bandwidth Estimation Tools. 11th IEEE Symposium on Computers and Communications, 2006, ISCC '06, pp. 709-718, 26-29 June 2006. Available at: http://pages.cs.wisc.edu/~pb/yaz_final.pdf. Last visited on: November, 2011.

[7] Pathchar tool websie. Available at: http://www.caida.org/tools/utilities/others/pathchar/. Last visited on: November, 2011.

[8] Clink tool website. Available at: http://allendowney.com/research/clink/. Last visited on: November, 2011.

[9] Trout tool website. Available at: http://allendowney.com/research/trout/. Last visited on: November, 2011.

[10] ICIR website: Tools for Bandwidth Estimation. Available at: http://www.icir.org/models/tools.html. Last visited on: November, 2011.

Depends on the type of work agreed on. In general, 30% investigation, 30% design, [20% implementation], 20% documentation
Network knowledge, some programming knowledge in C or/and Java, and basic linux knowledge

Supervisors: Prof. Dr. Burkhard Stiller, Dr. Guilherme Sperb Machado

back to the main page