Menu
▾
▴
[proxyTools-users] statProxyParallel - new option (was:Re: [noCensorship] +380 proxies)
|
From: wayne <wa...@ny...> - 2003-05-08 11:06:36
|
> From: Hat <ha...@ny...> > To: noc...@fr... > Subject: [noCensorship] +380 proxies > > Updated list of some proxies posted here. Tested with the new wayne's > monster :) statProxyParallel.pl 4.25 Hey! Since when is 67KB a monster? :-) > 12.239.75.65 :80 P 9.0/191.6 > 12.30.116.2 :80 P 9.8/179.3 > 130.127.130.253 :8000 P 9.3/176.9 > 130.161.36.7 :80 P 9.3/184.7 > 130.207.165.19 :80 P 9.9/177.3 > 142.179.103.64 :80 P 33.9/103.6 > 152.158.74.197 :80 P T 9.7/177.8 Ahh ... you meant a bandwidth monster. :-) It's true. This parallel statProxy will use every bps you have to get it's tests all done. Those reference page times (latency/complete) are pretty incredible. I guess you must have exceeded your download bandwidth. The latency times indicate this too. One test you used (zero) downloads the whole of the web page at www.panix.com to get those times. That page is about 14KB and SPP requested it about 375 times. That's about 5MB in a few minutes. The other test you used (anonymity) doesn't seem to have passed for any of these proxies. That's probably because they couldn't make the request back to you - no bandwidth. You could have used the '-s' option which inserts a wait between proxy socket operations. You'd need to try increasing values until the incoming data wasn't exceeding your bandwidth. A better way (for places where there is firewall blocking) is to simply tell SP to ignore proxies on blocked ports with '-b', say. Since you mainly want the standard port proxies for their anonymity, another option for you is to first use test 13/anonymity by itself. That test involves minimal downloaded data, and shouldn't cause you any bandwidth problems. You might then follow up with test 0 on the ones that passed, to get an idea of their speed. ******************************* But I've added a new option that should be better all around. :-) Version 4.27 includes the '-m' option. When you use the new option, test 0 is implied (so you don't need to specify any tests if you only want test 0) and modified to do only HEAD requests to the reference server, and check for the right response in the headers returned. No full page times are printed, but you still get the latency time (the time from request to the first data returned). Try that. The bandwidth problem should be minimized. Please let me know if it works for you. perl statProxyParallel.pl -m -l <proxyListFile> or perl statProxyParallel.pl -m <proxy:port>,<proxy:port>,... or perl statProxyParallel.pl -m http://url.of.forum.about.proxies/ ******************************* The only remaining test which needs download bandwidth is 14/noncensoring (only part page is actually required, but the whole lot probably arrives because of the long thin pipes from there to the USA). When I get time, I will also make '-m' modify this test (but not imply it like test 0). Note that test 0 with this option is less reliable. The reliability of test 14 would be reduced even more than for test 0, I think. [...] I understand you don't need to separate out the ones on UAE/KSA blocked ports any more, so I did the same tests again for the ones of use to people in UAE/KSA etc (without '-m', of course). Note: this was done using -b 80,3128,8080 to ignore proxies on those ports from the file. That option had a bug which I just fixed. With any older version, you probably need to write: -b #80#3128#8080#, but get v4.27. statProxy v4.26 report from xxx: 130.127.130.253 :8000 P P 3.5/18.2 via:WebSTAR Proxy (3.0b) 194.48.126.189 :8000 P P 4.4/21.4 via:1.0 194.48.126.189:8000 connBack:194.48.127.105 200.32.75.90 :8000 P P 3.6/13.5 202.64.159.200 :8000 P P 13.9/21.4 206.102.88.17 :8000 P 8.5/11.2 206.126.48.58 :8000 T T ?/? 210.131.177.253 :8000 P F 6.1/20.4 via:1.0 pacific.kyd.co.jp:8000 (Squid/2.4.STABLE6) 212.135.1.51 :31280 P 3.4/15.8 212.198.251.66 :8000 P P 4.5/21.4 216.126.204.21 :8888 T P ?/? connBack:65.166.64.132 216.148.244.37 :8000 T T ?/? 63.204.188.78 :8000 P P 6.0/13.6 via:WebSTAR Proxy (3.0b) 64.108.32.30 :8801 T T ?/? 64.132.153.94 :8888 P P 3.4/15.8 64.175.30.158 :8000 P P 3.4/15.8 via:WebSTAR Proxy (3.0b) 64.242.223.111 :8000 P R 3.6/13.5 64.242.223.125 :8000 R R ?/? 66.11.194.133 :8000 P P 3.6/13.5 via:WebSTAR Proxy (3.0b) 66.119.33.134 :8000 P F 3.5/18.2 via:1.1 csia3prx08.marketscore.com (NGP Diatom vfc3), 1.0 csia3che04 (NetCache NetApp/5.2.1R1D6) connBack:66.119.33.135 66.119.33.166 :8000 P F 3.5/18.2 via:1.1 csia4prx09.marketscore.com (NGP Diatom vfc3), 1.0 csia4che01 (NetCache NetApp/5.2.1R1) connBack:66.119.33.170 66.19.33.12 :8000 T T ?/? 66.250.69.1 :8888 P F 3.5/18.2 via:1.0 proxy.filtercube.com:8080 (Squid/2.4.STABLE7) connBack:66.250.68.40 Wall clock time: 0.91 mins. > Regards, > Hat Thanks for testing it Hat. I overlooked this bandwidth problem; the old statProxy was slow enough that download bandwidth was never a limitation. It crossed my mind many times that it could become one, but I forgot when I released the parallel thingy. -- wa...@ny... http://proxytools.sourceforge.net/ |
