w07 download speed really bad right now

Discussion in 'General troubleshooting' started by serializer, Jan 4, 2012.

  1. First I want to say that I like Winhost because fast support and easy setup. However, there is a problem when reporting issues. Winhost has no real external place where they can verify user problems.

    In my case I am reporting a serious issue on download speed. I present a clear case that download speed from a server on w02 is fast while w07 is really slow. Then they want me to do a "ping" which is not related to download speed. They say that speed is the same but the problem is that they test from a single location. Tracert on both domains points to the exact same network path.

    So, presenting two urls for you to test speed on:


    Fast: http://millimet.w02.Winhost.com/
    Slow: http://paydevco.w07.Winhost.com/

    I am testing from Sweden, but I have asked other users to test from other countries (not US yet). The fast one is 10 times fast than the slow.

    Fast: 300kb/s
    Slow: 30kb/s

    Interesting is that Upload speed is fast, about 800kb/s. Even if it don't matter my ping is about 180ms.

    Please test the urls and write what you experience. If you are on w07 and experience the same problem - please report to support!
     
    Last edited by a moderator: Oct 14, 2015
  2. Last edited by a moderator: Oct 14, 2015
  3. But testing from a single location is the best way to rule out server issues.

    We have two backbone connections at the data center, so your perceived download speed can be related to network speed along one of the two connections.

    Meaning one could be slower than another at a given time. But that is a network (Internet) issue, not a server issue.

    We can't control (nor can you) which backbone connection you may come in on. All that routing is done while the request is in transit across the Internet. A specific location may tend to use one connection more than the other, but - especially for users outside the U.S. - you could be routed to either one in what seems like a random fashion.

    We continuously monitor the backbone connections and balance them so that one doesn't become more saturated that the other. So if one appears to be slower than the other at a given time, we have to suspect Internet latency, which is why they ask for a ping or traceroute - those will show where the slow down may be occurring.
     
  4. While this discussion could go on forever it is really a waste of time as one network is clearly underperforming.

    I have now got results from different countries, including 2 UK locations 2 US locations, 2 Swedish locations, 1 russian location. All gives the same result.

    As I mentioned, the traceroute takes the same route but the last ip is different.

    Something is wrong and I can bet that if you give this link to your friends they will all find the exact same result.

    It is fine that you "continuously monitor" but I am presenting a very clear case now. I am not looking for a general answer. There is something wrong here. In all test cases, by all test persons the "faster" link is 10 or more times faster. If you use a backbone that provides this result on so many tests then you should really consider using another backbone.
     
  5. I wanted to chime in that I'm seeing similar results to above. One fast link, one slow link. I ran tests from a 10Mbps data center connection as well as from another 10Mbps+ link. I ran the tests at least twice per connection with similar results. Attached is a screenshot of the download rates from our data center's 10Mbps circuit.
     

    Attached Files:

  6. While I appreciate the trouble you are going to to make your point, we don't see the discrepancy (see the comparison below from a server more than 5,000 miles from our data center - the difference between the servers is virtually nil - .002 seconds), and if we can't see a discrepancy, there isn't anything for us to "fix."

    [​IMG]

    Please do not have any more of your friends register in the forum to post their results to your test. It isn't helping you prove your point.

    I believe that you saw a discrepancy as you described. I am not doubting you. But when we test we do not see it. So I'm afraid that there is no action for us to take.
     

    Attached Files:

    • wh.png
      wh.png
      File size:
      12.9 KB
      Views:
      698
    Last edited: Oct 14, 2015

Share This Page