fastos/tcpdive

A TCP performance profiling tool.

tcpdive's Issues

提错了

tcpdive.sh unable to find .config located under /usr/src/kernels/..

I'm not sure what the implication is but when I run tcpdive.sh, I see the following message regarding a missing file:

$ ./tcpdive.sh -v -R -f 50.1.2.2:*-50.1.2.4:53
Checking "/lib/modules/2.6.32-431.17.1.el6.x86_64/build/.config" failed with error: No such file or directory
$

When I try to search for the file, I note 'build' is linked to a different location under which .config is located:

$ ls -l /lib/modules/2.6.32-431.17.1.el6.x86_64/build 
lrwxrwxrwx. 1 root root 51 Jan 26 17:33 /lib/modules/2.6.32-431.17.1.el6.x86_64/build -> ../../../usr/src/kernels/2.6.32-431.17.1.el6.x86_64

$ ls -l /usr/src/kernels/2.6.32-573.12.1.el6.x86_64/.config 
-rw-r--r--. 1 root root 107139 Dec 15 16:49 /usr/src/kernels/2.6.32-573.12.1.el6.x86_64/.config
$

So, my question is, does the above error message matter for tcpdive.sh to function? If yes, how do I fix it?

Thanks in advance,

Keshav

Does tcpdive plan to maintain and roll out new versions over time

Cannot find a directory

i installed it via rpm, but there is nothing in log expect "Probe begin..."

and i run directly, it info me "Checking "/lib/modules/2.6.32-431.17.1.el6.x86_64/build/.config" failed with error: No such file or directory"

i'm sure i installed the correct kernel. what the problem i have?

Thanks in advance.

Troy

Improve TCP performance using control theoretic approach

(1) Evolution of TCP congestion control algorithm at UC Berkeley

1st generation: TCP Tahoe (1988) and TCP Reno (1990) --> Compound TCP (Windows OS)

2nd generation: TCP Friendly Rate Control (2000) --> TCP Cubic (Linux OS)
FAST TCP based on optimization theory (2006) and adaptive utility-based TCP based on control theory (2006)

3rd generation: Explicit Rate Control Protocol (2002) which needs to modify TCP header
XCP (2002), RCP (2005) and API-RCP (2007)

(2) Improve TCP performance using control theoretic approach

adaptive utility-based TCP based on control theory (2006)

[HY06] Y. Hong and O.W.W. Yang, "Self-Tuning Utility-Based Controller for End-to-End Congestion in the Internet," Proceedings of IEEE International Conference on Broadband Communications, Networks and Systems (IEEE BROADNETS), San Jose, California, U.S.A, October 2006.

https://www.researchgate.net/publication/4287116_Self-Tuning_Utility-Based_Controller_for_End-to-End_Congestion_in_the_Internet

Yang Hong's answer to question on control system design has been recommended by peer researchers as the most-popular answer at ResearchGate

"What are trends in control theory and its applications in physical systems (from a research point of view)?"

https://www.researchgate.net/post/What_are_trends_in_control_theory_and_its_applications_in_physical_systems_from_a_research_point_of_view2

Yang Hong's social media popularity per article = Number of followers/Number of articles = 152/36 = 4 followers per article (similar to 4 goals per hockey game) = top 0.1% popularity for technical articles evaluated by peer researchers = the most popular Ottawa-based researcher

https://www.researchgate.net/profile/Yang_Hong4

(3) Real-world implementation and evaluation of adaptive utility-based TCP based on control theory

Source code of TCP Cubic at GitHub

https://github.com/torvalds/linux/blob/master/net/ipv4/tcp_cubic.c

Source code developed by Yang Hong for idQ library of HybridAuth (All the source codes are committed by inBay Tech's CTO)

https://github.com/inbaytech

[Provider] idQ has been merged to the stable release v2.9.0 of HybridAuth

hybridauth/hybridauth#747

If I develop c code for adaptive utility-based TCP based on control theory, I do NOT have resource for evaluating the performance of adaptive utility-based TCP.

If the team of TCPDive wants to improve TCP performance using control theoretic approach, we can work together to make it happen.