VDS Performance Testing

    I choose VDS for myself - I thought about the issue of comparing performance.

    The purpose of this article is an attempt to find a criterion by which one can compare VDS from various providers and choose objectively the most successful price / quality combination. Perhaps the methods described in the article are not reliable, but as a starting point, it is quite enough.


    To begin with, as I see this question, if anything, correct me, because I am rather guided by common sense, because I have no experience with such tests.

    On the one hand, there is, for example, the ubench benchmark, which measures the power in parrots quite well, albeit surrogately. It should be noted that in the case of VDS, all such testers are not entirely correct, but with sufficient statistics, the situation looms quite close to real.

    However, I thought to use another way. There is a CMS that generates some pages of the site n seconds. I wrote a short script that requests 10 identical pages from different hosts, and then processes the result. The most interesting thing here is that, personally, I’m not quite sure of the correctness and adequacy of the methodology, so I’ll give the script sources:

    The script disp.php is the tunnel for AJAX, I decided not to steam it with cross-domain AJAX :)

    <? print file_get_contents($_SERVER["QUERY_STRING"]) * 1000; ?>

    The script exec.php launches the main CMS script, emulating the call, measures the call time and displays it Actually, the most important script is located at test.osmio .ru / disp.html - all JavaScript code, open source. Personally, I see several potentially “slippery” places:

    <?

    $time = microtime(1);
    // эмуляция обычного HTTP запроса
    file_get_contents("http://{$_SERVER["HTTP_HOST"]}/АДРЕС");
    print microtime(1) - $time;

    ?>





    1. If one script has not yet completed, and a new request has already arrived, the server starts to load stronger, but the checkmark for synchronous requests will save you from this
    2. The script does not measure the total time of the request, although perhaps it should: all servers respond with different delays ... But I deliberately did not include this in the calculation, because channels between servers can go as you like, including through Europe - and it happens that two neighboring DCs in Msk drive traffic in a roundabout way.
    3. HTTP request in exec.php - despite the fact that it goes to the same host

    After receiving the required number of tests, I slightly process them, as far as there are enough memories of the matan, and deduce.

    The server configurations are as follows:
    • vds-dbbb.1gb.ru - 1GB.ru, 500MHz, 384MB, OpenVZ, Gentoo
    • cms.dis.dj - Infobox.ru, 333MHz, 512MB, HyperV, Debian5
    • 89.188.123.24 - Ruweb.net, 2000MHz, 1536MB, VDSManager, FreeBSD6

    I measured the following values:
    HostThe averageDeviationMinMax0one23fourfive67eight9
    vds-dbbb.1gb.ru15644 (28%)128282136138131282137134134133144128
    cms.dis.dj129460 (5%)123414621275129112791259123414621282126912621316
    89.188.123.2444115 (3%)417460451420452456445460417445420443

    I repeat once again that the methodology is not a fact that is adequate to reality, but I will nevertheless draw some conclusions:
    • 1GB is a good sausage, although the values ​​are mostly minimal
    • Infobox shamelessly slows down - the processor and a fundamentally different virtualization technology are weaker there, but not as much
    • Ruweb is very stable and between approximately 1GB and ISPServer

    Ubench test results:
    • Infobox
      Ubench CPU: 27692
      Ubench MEM: 19291
      Ubench AVG: 23491
    • 1GB
      Ubench CPU: 55762
      Ubench MEM: 44521
      Ubench AVG: 50141
    • Ruweb - the car is evil in numbers, but it should be more powerful, the declared characteristics are a
      cut above the rest Ubench CPU: 600191
      Ubench MEM: 269780
      Ubench AVG: 434985

    Conclusions - Infobox is not such a weak VDSka, even despite the lower frequency - it competes with ISPServer. However, the generation time is much longer. I believe this is due to software settings. It should be noted here that everywhere except 1GB there is an ISPManager with Apache, PHP5, etc. configured by default. It should also be noted that there is a definite correlation between the ubench values ​​and what my tester intended.

    UPD. ISPServer ended the test period, and because its result was not very good at a rather impressive price for hosting - I excluded it from the calculation.

    Also popular now: