Smartone-Vodafone HSDPA Network Test Results

As most readers know I conducted a joint test of Smartone-Vodafone’s HSDPA network last week; which I now found out it cost me HKD60.00 in data charges, as I had detailed in my post, Smartone-Vodafone HSDPA Network Speed Tested.

At the time Smartone-Vodafone technicians did not have the data correlated until today.
SMV Network Test with SpeedTest
These results are from running the SpeedTest iPhone application on each of the iPhones used in the test.

SMV Network Test with Websites
These two data chart were provided by Smartone-Vodafone, so it has the tendency to show Smartone-Vodafone network as the faster network.

I would not have averaged out the speeds of the test on site by site bases, as we are talking about different locations on Hong Kong island at different time of the day. The better way to look at these data is to average out the browsing speed on a per site, per location bases. Below is the chart showing the speed in number of seconds to completely load the respective web page.
SMV Location View
As you can see from the chart above. It is not that web browsing speed of the iPhone 3GS on the Smartone-Vodafone network is always faster than the iPhone 3GS on the 3 HK network. The important thing is that the browsing speed is much more consistent for the iPhone 3GS on the Smartone-Vodafone network.

Having a consistent speed is much more important than having a fast downlink or uplink. I hope that this will provide consumers in Hong Kong a close to real life comparison of the two mobile carriers’ 3G networks.

7 Comments

  1. Thank you for sharing your results. The results described in this post was only testing the iPhone on the SMV 3G network as compared to the 3 HK 3G network.In either case it is best to restrict the destination web site to ones in Hong Kong, or when using SpeedTest iPhone application or web site, one should limit to testing traffic between the device and a Hong Kong server.Testing access to any servers outside of Hong Kong involves way too many additional variables to confidently blame SMV for the poor perforance.

    1. I do pass by occasionally, usually when ther is a reason. Over the last few weeks my “Home broadband” has degraded to the point wher accessing your blog is almost impossible. No, I don’t want you to intervene! I know that you are a concerned and involved user with some technical knowledge as am I.Ist repsonse I got from SMVs new (copy PCCW) “press 1 to…” customer service system, was to try incriminate my computers. I gave them the details they asked, and tried to explain that I was getting up to 90% packet loss. They had no idea what I was talking about, but at least were polite. 2nd response was that I had to turn off the power to the router and reconnect.Ho-hum.I then mentioned Ericccson who handle the whole network, routing to London and other factors (The BBC had gone offline)Your comment about keeping it to Hong Kong sites is an irrelevant simplification. Sure, it does remove international variables, but it doesn’t explain why Wharf are getting a packet loss of 20 to 90% from SMV. Wharf isn’t the problem, neither is Flagtel, nor Telia. The problem is that the network is saturated.Yes, I agree somewhat, in that SMV appears to be the best of a bad bunch. 3’s network performance sucks. (Its good at 3am on a Monday night from remote locations I’m told.) PCCW have admitted that they refuse access to their hotspots if the number of connections is too high. A friend is paying HK$300 a month or so for a USB modem connection that is, to say the least, “spotty”OFTA appear to have disappeared. A search of their site gives no repsonse to the speed test initiative announced in March 2010. The only test result I have ever seen is the one on your blog. Care to share the URL of the test page?None of the operators sites I’ve searched have a public “fair use” policy for anything other than a mobile phone. No mention is made, that I can find, of using your phone as a modem, although that is child’s play in some cases, and is equivalent to a home broadband in the demands it could place on the sytem.All in all, I’m beginning to think that HK’s much vaunted superiority, not just on mobile data, is a smoke screen. No-one is really promising anything except that you might get a connection of some kind if the wind blows in the right direction, or not. Just another load of hype.I supect that the pronouncement on what you can really expect was generated by complaints”fatre the intial launch. But please don’t call it “the truth about…”becasue it isn’t. It’s just corporate bullshit. to justify why they aren’t performing. Note that there are no figures at all as to what you can reasonably expect. Their used to be, but that page seems to have gone south, perhaps because it is embarasing. No idea.Read the current publicity (and bear in mind that this is the same network that you are on) which states that the expected speeds should be:Download: 2 to 4 Mbits/sUpload: 500 Kbit to 2 Mbit/sI get about 3 Mbit download on a good day to HK, 1Mbit on bad days, but about half the 500 Kbit upload. The download varies a lot, the upload is consistently bad. That is what you would expect if you are as cynical as I am, max out the download at the cost of upload. No reson why they can’t make both the same for any connection. Say äround 3 Mbit up plus down. That would be fine for me. Why should they care which way it is going? But ….that isn’t what happens in RL. Bizar.Current data as of now, plus or minus the time it takes me to do it:http://www.speedtest.net/resul…Ooof! Its a bad day. But then there is a prevailing Northeast breeze :-/BilouTo be fair, I’ll reboot and post a clean test, that is if the packet loss lets me.Oh, and in case you are wondering, yes, I do have records to prove the packet loss problem. The last time performance was an issue was perhaps a year ago. No time to chack now. That did get fixed eventually, perhaps a month later, and, as a bonus, they cut off all the ping responses in their network so we can’t check them any more.As a result of my most recent problem they have put one part of their internal network’s ping reposnse back on. Yup, you guessed it, the packet loss is internal to their network. Email me if you want the gory details as graphs and speed tests.Keep going. You seem to be a pretty knowledgeable, competent and fair guy, and you do speak the language I guess, which I don’t., so you have a better chance of getting to the bottom of this mess.

    2. I do pass by occasionally, usually when ther is a reason. Over the last few weeks my “Home broadband” has degraded to the point wher accessing your blog is almost impossible. No, I don’t want you to intervene! I know that you are a concerned and involved user with some technical knowledge as am I.Ist repsonse I got from SMVs new (copy PCCW) “press 1 to…” customer service system, was to try incriminate my computers. I gave them the details they asked, and tried to explain that I was getting up to 90% packet loss. They had no idea what I was talking about, but at least were polite. 2nd response was that I had to turn off the power to the router and reconnect.Ho-hum.I then mentioned Ericccson who handle the whole network, routing to London and other factors (The BBC had gone offline)Your comment about keeping it to Hong Kong sites is an irrelevant simplification. Sure, it does remove international variables, but it doesn’t explain why Wharf are getting a packet loss of 20 to 90% from SMV. Wharf isn’t the problem, neither is Flagtel, nor Telia. The problem is that the network is saturated.Yes, I agree somewhat, in that SMV appears to be the best of a bad bunch. 3’s network performance sucks. (Its good at 3am on a Monday night from remote locations I’m told.) PCCW have admitted that they refuse access to their hotspots if the number of connections is too high. A friend is paying HK$300 a month or so for a USB modem connection that is, to say the least, “spotty”OFTA appear to have disappeared. A search of their site gives no repsonse to the speed test initiative announced in March 2010. The only test result I have ever seen is the one on your blog. Care to share the URL of the test page?None of the operators sites I’ve searched have a public “fair use” policy for anything other than a mobile phone. No mention is made, that I can find, of using your phone as a modem, although that is child’s play in some cases, and is equivalent to a home broadband in the demands it could place on the sytem.All in all, I’m beginning to think that HK’s much vaunted superiority, not just on mobile data, is a smoke screen. No-one is really promising anything except that you might get a connection of some kind if the wind blows in the right direction, or not. Just another load of hype.I supect that the pronouncement on what you can really expect was generated by complaints”fatre the intial launch. But please don’t call it “the truth about…”becasue it isn’t. It’s just corporate bullshit. to justify why they aren’t performing. Note that there are no figures at all as to what you can reasonably expect. Their used to be, but that page seems to have gone south, perhaps because it is embarasing. No idea.Read the current publicity (and bear in mind that this is the same network that you are on) which states that the expected speeds should be:Download: 2 to 4 Mbits/sUpload: 500 Kbit to 2 Mbit/sI get about 3 Mbit download on a good day to HK, 1Mbit on bad days, but about half the 500 Kbit upload. The download varies a lot, the upload is consistently bad. That is what you would expect if you are as cynical as I am, max out the download at the cost of upload. No reson why they can’t make both the same for any connection. Say äround 3 Mbit up plus down. That would be fine for me. Why should they care which way it is going? But ….that isn’t what happens in RL. Bizar.Current data as of now, plus or minus the time it takes me to do it:http://www.speedtest.net/resul…Ooof! Its a bad day. But then there is a prevailing Northeast breeze :-/BilouTo be fair, I’ll reboot and post a clean test, that is if the packet loss lets me.Oh, and in case you are wondering, yes, I do have records to prove the packet loss problem. The last time performance was an issue was perhaps a year ago. No time to chack now. That did get fixed eventually, perhaps a month later, and, as a bonus, they cut off all the ping responses in their network so we can’t check them any more.As a result of my most recent problem they have put one part of their internal network’s ping reposnse back on. Yup, you guessed it, the packet loss is internal to their network. Email me if you want the gory details as graphs and speed tests.Keep going. You seem to be a pretty knowledgeable, competent and fair guy, and you do speak the language I guess, which I don’t., so you have a better chance of getting to the bottom of this mess.

  2. Erm, quite recently, maybe 2 months ago, SMV changed the routing of their home page. Until then they used to use a file download from it (89 Mbyte) as a test. Now that the page is routed more or less directly to the mast, that test is useless. I mentioned that to their tech support who were unaware of the change. They changed their testing method accordingly and now use an external site. Not sure which one.If the test is to ascertain if your mast to mobile connection is good then it is valid. But you can't compare between 3 and SMV by loading their respective homepages. Better to test SMV to 3s page and vice versa, then adjust for the difference in page size.My load times for 3's page http://www.three.com.hk from SMV are about 4 seconds (for about 278kb) after emptying the browser cache manually). I keep my cache on a 100 Mbyte ramdisk which is very fast. A hard disk would be a lot slower. On relaod with the cache already loaded up, it does it in less than 2 seconds. very difficult to measure accurately The browser is Opera 9.64. There are however a lot of non-network related factors in there: The browser render rate, the server latency, connection inital latency (maybe 900ms to get a connection set up through the router, then it drops to 50 – 70 ms for on-going downloads on the same browser connection (s). Browser is set to a max of 8 connexions to any one server and 64 to all servers combined. Thats a variable as well.

    1. Thank you for sharing your results. The results described in this post was only testing the iPhone on the SMV 3G network as compared to the 3 HK 3G network.

      In either case it is best to restrict the destination web site to ones in Hong Kong, or when using SpeedTest iPhone application or web site, one should limit to testing traffic between the device and a Hong Kong server.

      Testing access to any servers outside of Hong Kong involves way too many additional variables to confidently blame SMV for the poor perforance.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.