Wireless B slower than Ethernet?

Why is my Wireless (802.11b) connection sustantially slower than my Ethernet
connection?

The configurations for the two machines are...
- wireless (802.11b) on Win98 box through USB 1.1
- ethernet connection on W2K box.

I tried downloading a 3.2Meg file [1] simultaneously. The ethernet box took
14sec. That translates to a throughput of 1.8Mbps (3.2m*8/14sec=1.8Mbps).
The wireless box took over a minute.

I thought...
-USB 1.1 runs at 12Mbps
-802.11b runs at 11Mbps
-my internet connection runs at 1.8Mbps

Then, the USB 1.1 AirLink adapter shouldn't be the bottleneck. Any clue?

-w

[1] http://www.tucows.com/preview/198107.html

Yeah, I read it at...

http://www.overclockercafe.com/Revie...ter/Actiontec_
pg4.html
It's an old 2002 though...

Anyways, I'm not getting anything close to 5Mbps. My speed test reveals
that the ethernet connection is ~1.8Mbps. And, the wireless USB is 6 times
slower!!! Something is not right!



"Ray Taylor" <(E-Mail Removed)> wrote in message
news:40848608$(E-Mail Removed)...
> 802.11b can theoratically reach speeds of 11Mbps. In reality, you will
only
> get half that, 5Mbps.
> There could also be factors such as interference from things like other
> users on the same channel, cordless phones and microwaves.
>
>
> Ray Taylor
>
> "wendi" <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote in message
> news:tEZgc.12914$dZ1.1856@fed1read04...
> > Why is my Wireless (802.11b) connection sustantially slower than my
> Ethernet
> > connection?
> >
> > The configurations for the two machines are...
> > - wireless (802.11b) on Win98 box through USB 1.1
> > - ethernet connection on W2K box.
> >
> > I tried downloading a 3.2Meg file [1] simultaneously. The ethernet box
> took
> > 14sec. That translates to a throughput of 1.8Mbps
> (3.2m*8/14sec=1.8Mbps).
> > The wireless box took over a minute.
> >
> > I thought...
> > -USB 1.1 runs at 12Mbps
> > -802.11b runs at 11Mbps
> > -my internet connection runs at 1.8Mbps
> >
> > Then, the USB 1.1 AirLink adapter shouldn't be the bottleneck. Any
clue?
> >
> > -w
> >
> > [1] http://www.tucows.com/preview/198107.html
> >
> >
>
>

In article <tEZgc.12914$dZ1.1856@fed1read04>,
wendi <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote:
:I thought...
: -USB 1.1 runs at 12Mbps
: -802.11b runs at 11Mbps

802.11b's signalling rate is 11 megabits per second. Many of those bits
are used for preambles, requests to send, pauses between frames,
beacons, and error correction. The maximum end-to-end througnput
you can get with 802.11b depends on whether you are using udp or TCP,
and for the TCP case would be in the range of 5 to 6 megabits per second.

That assumes the maximum signalling rate is being used, that the signal
is strong enough for the maximum rate -- and it assumes that there
is no radio interference.
--
Tenser, said the Tensor.
Tenser, said the Tensor.
Tension, apprehension,
And dissension have begun. -- Alfred Bester (tDM)

802.11b can theoratically reach speeds of 11Mbps. In reality, you will only
get half that, 5Mbps.
There could also be factors such as interference from things like other
users on the same channel, cordless phones and microwaves.


Ray Taylor

"wendi" <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote in message
news:tEZgc.12914$dZ1.1856@fed1read04...
> Why is my Wireless (802.11b) connection sustantially slower than my
Ethernet
> connection?
>
> The configurations for the two machines are...
> - wireless (802.11b) on Win98 box through USB 1.1
> - ethernet connection on W2K box.
>
> I tried downloading a 3.2Meg file [1] simultaneously. The ethernet box
took
> 14sec. That translates to a throughput of 1.8Mbps
(3.2m*8/14sec=1.8Mbps).
> The wireless box took over a minute.
>
> I thought...
> -USB 1.1 runs at 12Mbps
> -802.11b runs at 11Mbps
> -my internet connection runs at 1.8Mbps
>
> Then, the USB 1.1 AirLink adapter shouldn't be the bottleneck. Any clue?
>
> -w
>
> [1] http://www.tucows.com/preview/198107.html
>
>

On Mon, 19 Apr 2004 19:35:56 -0400, in alt.internet.wireless , "wendi"
<FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote:

>Anyways, I'm not getting anything close to 5Mbps. My speed test reveals
>that the ethernet connection is ~1.8Mbps.

You say you measured this by downloading a file simultaneously from the
internet, and you have a 1.8Mbps internet connection? Then it doesn't
matter how fast your ethernet card is, 1.8 is the best you get (and thats
what you saw). By the way, thats a very fast internet connection. Are you
on a company LAN connection? Are you sure you're performing a fair measure
if so, remember company bandwidth is shared.

And then of course, you're halving each machine's performance further, by
downloading simultaneously..... do one at a time!
Possibly in this case the bandwidth was totally absorbed by the ethernet
machine, which then relinquished bandwidth only when finished.

>And, the wireless USB is 6 times slower!!! Something is not right!

As for the wireless, I'd expect that to be about 3-10x slower I'm afraid.
Even with excellent signal strength you'd get no better than 4Mbps from an
11b card, and with poor strength the performance could easily drop to
0.5Mbps or worse.


--
Mark McIntyre
CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
CLC readme: <http://www.angelfire.com/ms3/bchambless0/welcome_to_clc.html>


----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---

Ray Taylor <(E-Mail Removed)> wrote:
> 802.11b can theoratically reach speeds of 11Mbps. In reality, you will only
> get half that, 5Mbps.
> There could also be factors such as interference from things like other
> users on the same channel, cordless phones and microwaves.

In practice, I can see quite good speeds.
For example, at 5.5Mbps, I see well over 80% throughput.
(I hope to see similar results with 11g after I replace the 10base2
section of my network.)

> .......1.8 is the best you get .... By the way, thats a very fast
internet connection.
>
Hm... The average residential speed in California is ~1.6Mbps. 1.8 is very
reasonable.

> And then of course, you're halving each machine's performance further, by
> downloading simultaneously..... do one at a time!
>
I did one at a time, and also at different times. The results are somewhat
consistent for my non-scientific test. I can see why you would prefer the
different time method as it makes the math easier (no sharing bandwidth). I
choose the simultaneous method because usage of the destination site varies
from time to time. I realized that both methods are not very reliable. I
should really do my test through a speedtest site.

> As for the wireless ... you'd get no better than 4Mbps...
>
I understand that I would not get anything more than 5Mbps[1] for wireless
USB. My point is the wireless USB should not be the bottleneck. The
internet connection (1.8Mbps) which is subtantially lower than the wireless
USB should be the bottleneck. Hence, the wireless USB should be able to
provide 1.8Mbps.

-w
[1] See my previous post for reference


"Mark McIntyre" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> On Mon, 19 Apr 2004 19:35:56 -0400, in alt.internet.wireless , "wendi"
> <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote:
>
> >Anyways, I'm not getting anything close to 5Mbps. My speed test reveals
> >that the ethernet connection is ~1.8Mbps.
>
> You say you measured this by downloading a file simultaneously from the
> internet, and you have a 1.8Mbps internet connection? Then it doesn't
> matter how fast your ethernet card is, 1.8 is the best you get (and thats
> what you saw). By the way, thats a very fast internet connection. Are you
> on a company LAN connection? Are you sure you're performing a fair measure
> if so, remember company bandwidth is shared.
>
> And then of course, you're halving each machine's performance further, by
> downloading simultaneously..... do one at a time!
> Possibly in this case the bandwidth was totally absorbed by the ethernet
> machine, which then relinquished bandwidth only when finished.
>
> >And, the wireless USB is 6 times slower!!! Something is not right!
>
> As for the wireless, I'd expect that to be about 3-10x slower I'm afraid.
> Even with excellent signal strength you'd get no better than 4Mbps from an
> 11b card, and with poor strength the performance could easily drop to
> 0.5Mbps or worse.
>
>
> --
> Mark McIntyre
> CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
> CLC readme: <http://www.angelfire.com/ms3/bchambless0/welcome_to_clc.html>
>
>
> ----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet
News==----
> http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000
Newsgroups
> ---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption
=---

wendi wrote:
>> .......1.8 is the best you get .... By the way, thats a very fast
>> internet connection.
>>
> Hm... The average residential speed in California is ~1.6Mbps. 1.8
> is very reasonable.
>
>> And then of course, you're halving each machine's performance
>> further, by downloading simultaneously..... do one at a time!
>>
> I did one at a time, and also at different times. The results are
> somewhat consistent for my non-scientific test. I can see why you
> would prefer the different time method as it makes the math easier
> (no sharing bandwidth). I choose the simultaneous method because
> usage of the destination site varies from time to time. I realized
> that both methods are not very reliable. I should really do my test
> through a speedtest site.
>
>> As for the wireless ... you'd get no better than 4Mbps...
>>
> I understand that I would not get anything more than 5Mbps[1] for
> wireless USB. My point is the wireless USB should not be the
> bottleneck. The internet connection (1.8Mbps) which is subtantially
> lower than the wireless USB should be the bottleneck. Hence, the
> wireless USB should be able to provide 1.8Mbps.
>
> -w
> [1] See my previous post for reference
>
>
> "Mark McIntyre" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> On Mon, 19 Apr 2004 19:35:56 -0400, in alt.internet.wireless ,
>> "wendi" <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote:
>>
>>> Anyways, I'm not getting anything close to 5Mbps. My speed test
>>> reveals that the ethernet connection is ~1.8Mbps.
>>
>> You say you measured this by downloading a file simultaneously from
>> the internet, and you have a 1.8Mbps internet connection? Then it
>> doesn't matter how fast your ethernet card is, 1.8 is the best you
>> get (and thats what you saw). By the way, thats a very fast
>> internet connection. Are you on a company LAN connection? Are you
>> sure you're performing a fair measure if so, remember company
>> bandwidth is shared.
>>
>> And then of course, you're halving each machine's performance
>> further, by downloading simultaneously..... do one at a time!
>> Possibly in this case the bandwidth was totally absorbed by the
>> ethernet machine, which then relinquished bandwidth only when
>> finished.
>>
>>> And, the wireless USB is 6 times slower!!! Something is not right!
>>
>> As for the wireless, I'd expect that to be about 3-10x slower I'm
>> afraid. Even with excellent signal strength you'd get no better than
>> 4Mbps from an 11b card, and with poor strength the performance could
>> easily drop to
>> 0.5Mbps or worse.
>>
>>
>> --
>> Mark McIntyre
>> CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
>> CLC readme:
>> <http://www.angelfire.com/ms3/bchambless0/welcome_to_clc.html>
>>
>>
>> ----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet
>> News==---- http://www.newsfeed.com The #1 Newsgroup Service in the
>> World! >100,000 Newsgroups ---= 19 East/West-Coast Specialized
>> Servers - Total Privacy via Encryption =---

Some sites will give priority download speed to the first download from
a given IP address with the second download slowed. This prevents
bandwidth bandits from stealing their show. One test I use is to ftp
download from file storage on my ISP, and late at night local time. I
get the most consistent download estimates this way.

Q

wendi <FreeOfSpam_wendikun@no_spam_hotmail.com> wrote:
> I tried downloading a 3.2Meg file [1] simultaneously. The ethernet box took
> 14sec. That translates to a throughput of 1.8Mbps (3.2m*8/14sec=1.8Mbps).
> The wireless box took over a minute.

If you have access to a Unix or Windows command line somewhere on the wired
network, you can run iperf to get some readings on the speed. This will
still share a connection with other uses, but the reports are very stable
and reasonable. From my Unix ISP shell, I run
iperf -s -w 256k ( or some other reasonable window size )
on my Windows box I run
iperf -c ISP-unix-ip-address
This shows speed from the client to the server.
I can't reverse those roles because of NAT firewall.
If I run between two systems inside my network, I can use iperf in either
direction, or even test both ways simultaneously with -d
iperf -c server-ip-address -d
The bidirectional only works on some server/client pairs.

iperf might allow you to nail down where a problem resides.

http://dast.nlanr.net/Projects/Iperf/

---
Clarence A Dold - Hidden Valley (Lake County) CA USA 38.8-122.5

Just jumping into this thread so might be unrelated but...

My broadband connection regularly gets me more than 7.5mbit/second
(950Kbytes/second). Definately faster than a wireless B router would
provide.


"wendi" <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote in message
news:qIohc.17950$dZ1.12299@fed1read04...
> > .......1.8 is the best you get .... By the way, thats a very fast
> internet connection.
> >
> Hm... The average residential speed in California is ~1.6Mbps. 1.8 is
very
> reasonable.
>
> > And then of course, you're halving each machine's performance further,
by
> > downloading simultaneously..... do one at a time!
> >
> I did one at a time, and also at different times. The results are
somewhat
> consistent for my non-scientific test. I can see why you would prefer the
> different time method as it makes the math easier (no sharing bandwidth).
I
> choose the simultaneous method because usage of the destination site
varies
> from time to time. I realized that both methods are not very reliable. I
> should really do my test through a speedtest site.
>
> > As for the wireless ... you'd get no better than 4Mbps...
> >
> I understand that I would not get anything more than 5Mbps[1] for wireless
> USB. My point is the wireless USB should not be the bottleneck. The
> internet connection (1.8Mbps) which is subtantially lower than the
wireless
> USB should be the bottleneck. Hence, the wireless USB should be able to
> provide 1.8Mbps.
>
> -w
> [1] See my previous post for reference
>
>
> "Mark McIntyre" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
> > On Mon, 19 Apr 2004 19:35:56 -0400, in alt.internet.wireless , "wendi"
> > <FreeOfSpam_wendikun@NO_SPAM_hotmail.com> wrote:
> >
> > >Anyways, I'm not getting anything close to 5Mbps. My speed test
reveals
> > >that the ethernet connection is ~1.8Mbps.
> >
> > You say you measured this by downloading a file simultaneously from the
> > internet, and you have a 1.8Mbps internet connection? Then it doesn't
> > matter how fast your ethernet card is, 1.8 is the best you get (and
thats
> > what you saw). By the way, thats a very fast internet connection. Are
you
> > on a company LAN connection? Are you sure you're performing a fair
measure
> > if so, remember company bandwidth is shared.
> >
> > And then of course, you're halving each machine's performance further,
by
> > downloading simultaneously..... do one at a time!
> > Possibly in this case the bandwidth was totally absorbed by the ethernet
> > machine, which then relinquished bandwidth only when finished.
> >
> > >And, the wireless USB is 6 times slower!!! Something is not right!
> >
> > As for the wireless, I'd expect that to be about 3-10x slower I'm
afraid.
> > Even with excellent signal strength you'd get no better than 4Mbps from
an
> > 11b card, and with poor strength the performance could easily drop to
> > 0.5Mbps or worse.
> >
> >
> > --
> > Mark McIntyre
> > CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
> > CLC readme:
<http://www.angelfire.com/ms3/bchambless0/welcome_to_clc.html>
> >
> >
> > ----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet
> News==----
> > http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000
> Newsgroups
> > ---= 19 East/West-Coast Specialized Servers - Total Privacy via
Encryption
> =---
>
>