Selfmade Private vs. Bought Private vs. Semi-dedicated Proxies
Based on the two previous tests here and here I decided to do one more take on the proxy testing issue and which proxies were more reliable. In this test I tested private proxies I set up myself (called: mpp), private proxies I bought from a well-known provider (called: bpp) and semi-dedicated proxies also gotten from a well-known provider (called: semi).
The setting: about 3 hours of running a list with 5 mpp, 5bpp and 10 semi - first at 100 threads (5 thread per proxy) then about 3 hours at 200 threads (10 threads per proxy). Recaptcha resolve errors were removed prior to checking the results.
Results:
mpp - 100 threads = 0.61 fail message participation per minute
mpp - 200 threads = 0.89 fail message participation per minute
increase in fail message participation: 46%
bpp - 100 threads = 1.17 fail message participation per minute
bpp - 200 threads = 1.47 fail message participation per minute
increase in fail message participation: 26%
semi - 100 threads = 1.37 fail message participation per minute
semi - 200 threads = 2.07 fail message participation per minute
increase in fail message participation: 51%
Conclusion:
As in the previous tests self-made private proxies are far better than bought ones, followed by semi-dedicated proxies. The interesting thing is that operating semi at 5 threads per proxy seems to have more or less the same reliability as operating bpp at 10 threads, with the price being more or less the same ($1 per semi * 5 threads = 5 threads per $; $2 per bpp * 10 threads = 5 threads per $) semi dedicated proxies give you a wider scope of ip addresses for the same price as bpp (with a slight + for semi).
It is to be noted however that 1.37 fail message participation per minute (semi at 5 threads) means 1973 fail message participation per day while running semi at 10 threads per proxy means an average of 2981 fail message participation per day. It does not mean the proxies failed that many times, it just means that whatever went wrong increasing the load on your proxies will cause them to fail more regardless of other errors and causes.
Whether or not the price of proxies is worth it for you, however, remains something of personal preference.
What do you think? Comments, suggestions?
The setting: about 3 hours of running a list with 5 mpp, 5bpp and 10 semi - first at 100 threads (5 thread per proxy) then about 3 hours at 200 threads (10 threads per proxy). Recaptcha resolve errors were removed prior to checking the results.
Results:
mpp - 100 threads = 0.61 fail message participation per minute
mpp - 200 threads = 0.89 fail message participation per minute
increase in fail message participation: 46%
bpp - 100 threads = 1.17 fail message participation per minute
bpp - 200 threads = 1.47 fail message participation per minute
increase in fail message participation: 26%
semi - 100 threads = 1.37 fail message participation per minute
semi - 200 threads = 2.07 fail message participation per minute
increase in fail message participation: 51%
Conclusion:
As in the previous tests self-made private proxies are far better than bought ones, followed by semi-dedicated proxies. The interesting thing is that operating semi at 5 threads per proxy seems to have more or less the same reliability as operating bpp at 10 threads, with the price being more or less the same ($1 per semi * 5 threads = 5 threads per $; $2 per bpp * 10 threads = 5 threads per $) semi dedicated proxies give you a wider scope of ip addresses for the same price as bpp (with a slight + for semi).
It is to be noted however that 1.37 fail message participation per minute (semi at 5 threads) means 1973 fail message participation per day while running semi at 10 threads per proxy means an average of 2981 fail message participation per day. It does not mean the proxies failed that many times, it just means that whatever went wrong increasing the load on your proxies will cause them to fail more regardless of other errors and causes.
Whether or not the price of proxies is worth it for you, however, remains something of personal preference.
What do you think? Comments, suggestions?
Comments
I would be interested to know if the self made deteriorate with age though.