DSL SPEED

I have a ADSL modem rated at 1.5mps

A test from an internet site computed the DSL speed .

798 kbps download 310 kbps upload Quality of service 79%

Is the above good service ? Should the download be about 1.5mps

What should I expect from the upgrade to 3.0 MPS

That's way too low. I have ADSL service from Verizon which is advertised as 1.5/384 and syncs at 1792/448k bps (which is known as over provisioning) and receive 1525k bps downloads/370k bps uploads.

Most other DSL providers do NOT over provision, so if you're getting

80-85% or more of your package speed, you're doing good. The overhead is about 15%.

Now to the core of the issue: web based tests are often quite inaccurate. Where are you located geographically? If you're going to rely on a web test, make sure it's close to you geographically. Better yet, download a file via FTP from any fast public server - Here are a few that give me steady topped out speeds and are load balanced [I'm in NY btw]:

ftp://ftp.ea.com/pub/easports/demos/madden2004/Madden2004_Demo.exe ftp://ftp.aol.com/aol6.0/171340/setupaol60.exe ftp://ftp1.optonline.net/test64

Also if you're going to be using web based tests, use the DSLReports JAVA tests at the nearest Speakeasy POP:

sfo.speakeasy.net San Francisco lax.speakeasy.net Los Angeles dfw.speakeasy.net Dallas sea.speakeasy.net Seattle nyc.speakeasy.net New York blt.speakeasy.net Baltimore

You'll find many more here:

(just remember, JAVA tests are no where near accurate as distance plays a huge role and so are overloaded test servers.)

People who have 3.0/768 are reporting ~2800k bps downstream and ~720k bps upstream on Verizon ADSL.

If your modem has difficulty maintaining sync at 1.5 or have borderline signal qualities, 3.0 is more demanding (clean line, good snr) and may push your modem to loose sync. You can't know until you try. :)

HTH Don't trust Web Based Tests is the point of this post.

FWIW on this cable connection, web tests rate it anywhere from 1m bit to java tests at 8m bps or so. Here's it's true speed:

% docsdiag -v | grep -i "qos" QoS max upstream bandwidth = 1000000 bps QoS max downstream bandwidth = 10000000 bps

% wget ftp://ftp.ea.com/pub/easports/demos/madden2004/Madden2004_Demo.exe

--10:51:50-- ftp://ftp.ea.com/pub/easports/demos/madden2004/Madden2004_Demo.exe => `Madden2004_Demo.exe' Resolving ftp.ea.com... 159.153.229.107, 159.153.229.106 Connecting to ftp.ea.com[159.153.229.107]:21... connected. Logging in as anonymous ... Logged in! ==> SYST ... done. ==> PWD ... done. ==> TYPE I ... done. ==> CWD /pub/easports/demos/madden2004 ... done. ==> PORT ... done. ==> RETR Madden2004_Demo.exe ... done. Length: 69,754,368 (unauthoritative)

100%[====================================>] 69,754,368 1.15M/s ETA 00:00

10:52:52 (1.07 MB/s) - `Madden2004_Demo.exe' saved [69754368]

HTH Karyn Lightspeed Underground

try posting to COMP.DCOM.XDSL

If you upgrade to 3.0, you will get: ~798 kbps download ~310 kbps upload

if you test from the same testing site under the same conditions.

You should run speed tests from several locations, and the FASTEST speed test is closest to your actual service speed. That's because speed tests depend on the speed test server and the Internet, as well as your local connection to it. So your speed test results are only going to be as fast as (your connection, or the speed test server, or the connections between then, which ever is slowest)

But assuming that you ran several speed tests, and the best download speed you got was 798kbps, then you could conclude that your particular ADSL connection is not as fast as it theoretically COULD be, due to local line conditions and/or distance from CO. Therefore, upgrading to 3.0 would not gain you anything.

Find your state, or a nearby state, on the following list. Run several tests. If the fastest result you get is over 1500kbps, then you might want to consider upgrading to 3.0. If you never even HIT 1500kbps running tests from several servers in your state or nearby, then your connection is already as fast as it's going to be. -Dave