Airport Express/AE card strangeness

[Chip NoVaMac]

In trying to decide what I was going to do if/when I got a Mac mini and connecting to the internet, I decided to do a test with my PB and eMac. Wasn't sure that I wanted to do a BTO to get the AE installed.

I have Cox cable modem service distributed via Airport Express. I setup the eMac as the source via the ethernet port. And of course connected the PB as the receiver of that signal via the ethernet port. I did turn off the Airport from the menu bar.

The strangeness is that when testing MR links in Safari, the ethernet option is opening pages faster. Much faster than with using the Airport card in the PB. Did do a broadband speed test, and the Emac gets about 3.3mbit average via the Airport. The same for the PB with Airport or ethernet from the eMac. Any reason for this?

 

[jsw]

Not sure I'm following correctly....

My understanding is that you have an AirPort Express connected to your cable modem, and that the eMac is receiving the signal wirelessly (via AirPort), then using its ethernet port to connect to the PB.

Is this correct?

Are all AirPort cards AirPort Extreme cards?

 

[Chip NoVaMac]

Not sure I'm following correctly....

My understanding is that you have an AirPort Express connected to your cable modem, and that the eMac is receiving the signal wirelessly (via AirPort), then using its ethernet port to connect to the PB.

Is this correct?

Are all AirPort cards AirPort Extreme cards?

Right on all accounts.

When I run the Airport on the PB, it takes a bit longer for the pages to start to download vs. by running it from the eMac via the ethernet. The PB does show one less bar, and they are in the same room.

 

[jsw]

Right on all accounts.

When I run the Airport on the PB, it takes a bit longer for the pages to start to download vs. by running it from the eMac via the ethernet. The PB does show one less bar, and they are in the same room.

That's your answer - the PBs have notoriously bad reception, and the lower signal strength probably drops it to 802.11b speeds. A mini should receive better - the antenna, I think, doesn't have to go through metal but is instead near the plastic top. Reception should be much better. Not as good as the eMac, but better than the PB.

Edit: probably essentially as good as the eMac.

 

[Chip NoVaMac]

Thanks for the help JSW, but I do have the Airport Express setup to "send" only in "g" not "b". But yet broadband speed tests show that on the PB both connections are about the same. It is about the slowness of web-pages opening or connecting via Airport compared to ethernet with the PB. I am just stumped.

 

[jsw]

Thanks for the help JSW, but I do have the Airport Express setup to "send" only in "g" not "b". But yet broadband speed tests show that on the PB both connections are about the same. It is about the slowness of web-pages opening or connecting via Airport compared to ethernet with the PB. I am just stumped.

Odd, that. I'm stumped too. Encryption/NAT messing with setting up connections? Speed checks send one big file, so they'd get hit once, whereas typical web pages create multiple connections when loading.....

FWIW, 802.11g will downgrade speeds to 802.11b levels or slower if the signal strength lessens. It just isn't technically 802.11b, but a slower, 802.11b-like speed.

 

[Chip NoVaMac]

Odd, that. I'm stumped too. Encryption/NAT messing with setting up connections? Speed checks send one big file, so they'd get hit once, whereas typical web pages create multiple connections when loading.....

FWIW, 802.11g will downgrade speeds to 802.11b levels or slower if the signal strength lessens. It just isn't technically 802.11b, but a slower, 802.11b-like speed.

Thanks for the further help. The speed tests I used were from http://www.dslreports.com (http://www.dslreports.com/stest?loc=97). They seem to use a variety of file sizes to do the test.

 

[robbieduncan]

I see this too. I've always put it down to 802.11 (all falvours) having much higher latency than Ethernet. When you are browsing the web most files that you download are very small so latency has a much bigger effect than overall throughput. Try pinging your basestation (or the cable modem/router is probably a better idea) both wirelessly and wired. For me the differences are massive:

Wireless:
ping 192.168.0.1
PING 192.168.0.1 (192.168.0.1): 56 data bytes
64 bytes from 192.168.0.1: icmp_seq=0 ttl=64 time=1.992 ms
64 bytes from 192.168.0.1: icmp_seq=1 ttl=64 time=1.844 ms
64 bytes from 192.168.0.1: icmp_seq=2 ttl=64 time=4.425 ms

Wired:
ping 192.168.0.1
PING 192.168.0.1 (192.168.0.1) 56(84) bytes of data.
64 bytes from 192.168.0.1: icmp_seq=1 ttl=64 time=0.603 ms
64 bytes from 192.168.0.1: icmp_seq=2 ttl=64 time=0.574 ms
64 bytes from 192.168.0.1: icmp_seq=3 ttl=64 time=0.652 ms


Also note that the wired times show much less variation.

 

[Chip NoVaMac]

I see this too. I've always put it down to 802.11 (all falvours) having much higher latency than Ethernet. When you are browsing the web most files that you download are very small so latency has a much bigger effect than overall throughput. Try pinging your basestation (or the cable modem/router is probably a better idea) both wirelessly and wired. For me the differences are massive:

Wireless:
ping 192.168.0.1
PING 192.168.0.1 (192.168.0.1): 56 data bytes
64 bytes from 192.168.0.1: icmp_seq=0 ttl=64 time=1.992 ms
64 bytes from 192.168.0.1: icmp_seq=1 ttl=64 time=1.844 ms
64 bytes from 192.168.0.1: icmp_seq=2 ttl=64 time=4.425 ms

Wired:
ping 192.168.0.1
PING 192.168.0.1 (192.168.0.1) 56(84) bytes of data.
64 bytes from 192.168.0.1: icmp_seq=1 ttl=64 time=0.603 ms
64 bytes from 192.168.0.1: icmp_seq=2 ttl=64 time=0.574 ms
64 bytes from 192.168.0.1: icmp_seq=3 ttl=64 time=0.652 ms


Also note that the wired times show much less variation.

Sorry, maybe I did not explain it better; but your post looks like in the right track for the answer. But the "wired" PB is working off a wireless connection from the eMac, so should not the latency be the same?

 

[robbieduncan]

Sorry, maybe I did not explain it better; but your post looks like in the right track for the answer. But the "wired" PB is working off a wireless connection from the eMac, so should not the latency be the same?

Not really. There is latency on every leg of the connection. So there is some latency between the PB and the eMac as well as any latency between the eMac and the the outside world. In addition there will be additional latency for the eMac to route packets from the wireless to the wired connection. You can measure all this via ping.

To me it looks like the eMac routing is introducing most of the latency here. Is it running something that is eating all the CPU time (Folding, SETI or something)?

 

[HeWhoSpitsFire]

It is not your wireless b.

Ethernet is either 1000/base or 100/base TX. Both move faster than wireless (802.11g) , not to mention 802.11b is is up to 11Mbps, while 100/base TX us 100Mbps.

802.11g provides speeds up to 54Mbps.

Most current Macs have Ethernet support 100/base TX or 1000/base. Both protocols surpass wireless substantially.

"Combined Latency" sounds like your culprit.

Not that anyone is short of this info, but it sounds like the issue. No matter the setup, the speeds will always differ, no matter the outside connection.

Most Ethernet hookups move clos to 0.4 or .6, my gig eto1(1000/base to 1000/base is less than .15).

I really hope none of you still have 10/base T hookups!

(All speeds are burst transfer rated. Does not illustrate true throughput speeds.)
edited for: sp