1. chuvisco's Avatar
    This is probably my first post here, but I use the same username at HF ...

    I just ran two speed tests using the xtremelabs app on a Tour at the VZW store in Midtown Atlanta (Ponce de Leon), and the best result was 313 kbps down, 278 kbps up. The handset was indicating 5 bars EV. Needless to say, this is pretty disappointing for a Rev. A device, but it mirrors my experience with my 8330.

    On the 8330, I've run a lot of tests, always with strong EVDO signal, both tethered (speedtest.net) and not (xtremelabs app), and the best I've gotten was 246 kbps:

    Jun 25 23:38:00

    Here are the Speedtest results for my Blackberry:

    Connection Type: Carrier
    Avg. Download Speed: 226.13Kbits/s
    Max. Download Speed: 246.39Kbits/s

    Avg. Upload Speed: 33.52Kbits/s
    Max. Upload Speed: 58.13Kbits/s
    I've also noted that it often takes a minute or more to download a ~1MByte (8 Mbit) email attachment, corresponding to a bandwidth of ~133 kbps. The 8330 is Rev. 0, but this hardly matches the 400-700 kbps generally quoted even for that tech.

    I've seen a handful of decent speed test results from the Storm/8330, but IIRC, the majority are closer to the results above than Verizon's quoted EVDO speeds. Most of the Bold speed test results I've seen are respectable, so I don't think the BB servers are the bottleneck. Do you all see similar speeds, and does anyone know why I'm getting such poor performance on multiple devices in multiple locations across the Atlanta area?

    I have links, but CB won't let me post them ...
    07-13-09 03:23 PM
  2. chuvisco's Avatar
    For the Tour on Rev. A, according to Verizon, "typical" speeds should be 600 kbps to 1.4 Mbps. Is anyone getting this?

    You can find the speedtest app at:

    xtremelabs.com/apps
    07-13-09 05:06 PM
  3. ericsmcdonald's Avatar
    This is probably my first post here, but I use the same username at HF ...

    I just ran two speed tests using the xtremelabs app on a Tour at the VZW store in Midtown Atlanta (Ponce de Leon), and the best result was 313 kbps down, 278 kbps up. The handset was indicating 5 bars EV. Needless to say, this is pretty disappointing for a Rev. A device, but it mirrors my experience with my 8330.

    On the 8330, I've run a lot of tests, always with strong EVDO signal, both tethered (speedtest.net) and not (xtremelabs app), and the best I've gotten was 246 kbps:
    I think Twins can give the definitive on this one, but the Rev. A is going to be most notable when you tether. When you're just using your Blackberry, you're going through the RIM servers. They will slow you down.

    If you tether your device to a laptop and do a speed test, you'll notice blazing fast speeds. When you use your BlackBerry device, it goes through RIM. When you use your device as a modem, it goes straight to the internet.
    07-13-09 05:10 PM
  4. TwinsX2Dad's Avatar
    First off, speed tests through the BlackBerry browser are inaccurate, as they measure the speed of the connection to the RIM servers. If using OperaMini, you compound the issue by now running through two intermediate servers.

    What are you using to tether? If Tetherberry or another proxy, you have yet another layer, plus carriers are now beginning to detect & throttle back these connections.

    I always recommend tethering a decent computer, using WinDUN. Access Manager adds its own issues & does slow things down.

    I was in Atlanta about 8 weeks ago and found speeds to be well above average.

    Posted from my CrackBerry at wapforums.crackberry.com
    07-13-09 05:16 PM
  5. greens601's Avatar
    i love atlanta (not the braves)
    07-13-09 05:34 PM
  6. chuvisco's Avatar
    First off, speed tests through the BlackBerry browser are inaccurate, as they measure the speed of the connection to the RIM servers. If using OperaMini, you compound the issue by now running through two intermediate servers.

    What are you using to tether? If Tetherberry or another proxy, you have yet another layer, plus carriers are now beginning to detect & throttle back these connections.

    I always recommend tethering a decent computer, using WinDUN. Access Manager adds its own issues & does slow things down.

    I was in Atlanta about 8 weeks ago and found speeds to be well above average.

    Posted from my CrackBerry at wapforums.crackberry.com
    Thanks for the replies! My understanding is that, when using the socket connection in Opera Mini, it connects directly to the Opera servers (my 8330 asks me to approve a low-level network connection). I would assume that Xtreme Labs would do the same in their speed test app, but I don't know for sure. It doesn't make the same request for a low-level connection, but maybe their app is signed vs. the unsigned Opera Mini?

    In any case, one of the real-world use cases that matters to me is how long it takes to download an attachment, and I'm not really satisfied with a throughput of 1-200 kbps on a network that should support 2-3 times that. At Rev. 0 speeds (say 400 kbps), a 1 MByte attachment should download in 20 sec. At Rev. A speeds (say 600 kbps), it should download in 13 sec.

    I was skeptical of the low number reported by the Xtreme Labs app at first, but after comparing it to my email attachment download rate, it seemed about right. I would be interested in seeing numbers from others--attachment download times (it needs to be a big attachment, say ~1 MB or more, so that we are measuring download speed and not latency) or results from the speed test app.

    I'm a little frustrated with my BB experience because there are mobile devices out there that download files and return speed test results at more or less the specified network speeds ...
    07-13-09 06:23 PM
  7. sprke81's Avatar
    In any case, one of the real-world use cases that matters to me is how long it takes to download an attachment, and I'm not really satisfied with a throughput of 1-200 kbps on a network that should support 2-3 times that. At Rev. 0 speeds (say 400 kbps), a 1 MByte attachment should download in 20 sec. At Rev. A speeds (say 600 kbps), it should download in 13 sec.



    I'm a little frustrated with my BB experience because there are mobile devices out there that download files and return speed test results at more or less the specified network speeds ...

    Thats going to be the way blackberry is. Its not because evdo is slow. If you want speeds that are more like the true network speeds then you'll have to go with a different os platform.
    07-13-09 06:53 PM
  8. mpafr012's Avatar
    Twins, could the reason also be that the stores are not using a full cellular site by something like a network extender? The reason I ask is I always have 5 bars in a store and not always outside of a store?
    07-13-09 07:01 PM
  9. sprke81's Avatar
    Twins, could the reason also be that the stores are not using a full cellular site by something like a network extender? The reason I ask is I always have 5 bars in a store and not always outside of a store?
    I know that I'm not twins, but yea they have signal boosters in the stores.
    I"m a electrician and I wired a VZ store last year that was right across the road from a VZ tower. It didn't matter we still had to install a signal booster. Its a white dome looking thing in the ceiling.

    So that might affect your speeds at some stores but I really kinda doubt it.
    07-13-09 07:06 PM
  10. gvillager's Avatar
    07-13-09 07:24 PM
  11. TwinsX2Dad's Avatar
    You're not far off on your assessment of what happens. BlackBerrys can be difficult to get an accurate speed test on AND they are much less able to utilize the speed available from the network - all because of the intermediate connection needs. We see much higher real-world download speeds, regardless of network, using devices other than BlackBerry.

    Using the BlackBerry browser, everything has to hop through the RIM servers. As we've discussed, Opera can add another layer. If I use a Palm or WinMo device, I get rid of those hops & end up with a faster connection - normally because the Internet servers are closer. RIM doesn't maintain multiple server locations for any region, so those of us farther away get stiffed.

    If you use tethering & WinDUN, you're using only the carrier network & the closest available server without any additional packet monitors.

    As I sit here, Tour & Saga on my desk - both rev a - just for fun, I downloaded the exact same file from the same server (ours). BlackBerry did it in a somewhat admirable 55 seconds. Saga took 23. Tried another file, then a third - this time radio show podcasts. Again, roughly 2:1 in favor of Saga. Grabbed a HTC Touch - similar results. Tried a Palm Pre on Sprint, same thing.

    BlackBerry seemed fast until I did a side-by-side.

    BlackBerry is the preferred platform of many, and I get that - we have opted for it in my business, because it is the best email platform & is the simplest to remotely administer & monitor. It is not the best for network connectivity or download speeds. The competition - including iPhone - blows it away, when compared on like networks. It shouldn't matter much when tethering, but your tethering percentage is something only you can assess.

    Posted from my CrackBerry at wapforums.crackberry.com
    07-13-09 08:18 PM
  12. chuvisco's Avatar
    This thread may provide some info:
    Yes, the 350 kbps max. mentioned there is in agreement with my experience in the Atlanta area.

    You're not far off on your assessment of what happens. BlackBerrys can be difficult to get an accurate speed test on AND they are much less able to utilize the speed available from the network - all because of the intermediate connection needs. We see much higher real-world download speeds, regardless of network, using devices other than BlackBerry.

    Using the BlackBerry browser, everything has to hop through the RIM servers. As we've discussed, Opera can add another layer. If I use a Palm or WinMo device, I get rid of those hops & end up with a faster connection - normally because the Internet servers are closer. RIM doesn't maintain multiple server locations for any region, so those of us farther away get stiffed.

    If you use tethering & WinDUN, you're using only the carrier network & the closest available server without any additional packet monitors.
    I was using TetherBerry, so I'll concede the tethering point, but the poor on-device speed still bugs me.

    As I sit here, Tour & Saga on my desk - both rev a - just for fun, I downloaded the exact same file from the same server (ours). BlackBerry did it in a somewhat admirable 55 seconds. Saga took 23. Tried another file, then a third - this time radio show podcasts. Again, roughly 2:1 in favor of Saga. Grabbed a HTC Touch - similar results. Tried a Palm Pre on Sprint, same thing.

    BlackBerry seemed fast until I did a side-by-side.

    BlackBerry is the preferred platform of many, and I get that - we have opted for it in my business, because it is the best email platform & is the simplest to remotely administer & monitor. It is not the best for network connectivity or download speeds. The competition - including iPhone - blows it away, when compared on like networks. It shouldn't matter much when tethering, but your tethering percentage is something only you can assess.

    Posted from my CrackBerry at wapforums.crackberry.com
    Thanks for the data points--this is interesting. I wish this difference between Blackberry and other platforms were more apparent to consumers in the smartphone market.

    Twins, do you (or anyone else) know if there is a difference in speed between a well-managed corporate BES and Verizon's BIS?
    07-13-09 09:06 PM
LINK TO POST COPIED TO CLIPBOARD