0

I have a router that spreads my internet signal using LAN and Wlan. I connected another wlan-router per LAN on the other side of my apartment. So 2 routers, connected by LAN-cable.

I configured the second router as access point and chose a unique WLAN-SSID. Was that clever? I read that the idea of an access point is, that the device of moving users always automatically allocates itself to the nearest spot. This is not possible if an access point doesn't share the same SSID, right? Or would 2 equal SSID's fight each other? Are there some other pitfalls like the channel they are using (wasn't able to configure that one)?

The cable that I plug into the second router gives me 200k+ mbit. If I plug it into the second router and use the WLAN network from it, I get like 40 mbit, while sitting directly next to it. Maybe the router is just weak, or something else is going on... But the router is clearly slowing down the signal tremendously. That's not just the usual Wlan issue (300mbit according to specification).

EDIT: Just checked the lan connection. Did a speedtest on my TV that is connected with router 2 by LAN. The speed is terrible: 20 mbit. The issue cannot be related to the wifi signal. I used this router as main router once, where I didn't had such problems. Could this be related to a badly programmed access point api?

1

Last thing first - "300 Mbit" writ large on the AP box is not all it's cracked up to be - step one, it's half-duplex; step two, it includes all the wireless overhead (traffic that has to be sent and received, but which is only to maintain the wireless link, not to move your data across it.) A wire, by contrast, is full-duplex and lacks the wireless overhead.

90 Mbit (as if you were connected to a cable) is about as much real throughput as you can expect from a "300 Mbit" AP under perfect conditions. You may have additional limitations depending on the network hardware in your laptop. You may also have limitations from other devices using the same AP or from interference from elsewhere. Wires basically always win...

Having done it both ways, having both APs use the exact same SSID and security settings is generally preferable. As for "fighting" each other, keep them on different channels - which in the 2.4 GHz band means, 1, 6, and 11 and only 20 MHz - at 40 MHz bandwidth, there are no non-interfering channels in the 2.4 band. The "channels" in that band are far narrower than the bandwidths used, so channels 2-5 interfere with 1&6, channels 7-10 interfere with 6&11. 5GHz generally has more available channels and those channels are spaced further apart, so there is less chance of interference.

Right next to the router is not always the fastest place to sit. If the signal is too strong, the input stage of the radio may be overloaded and that will actually make it slower than a more ideal signal level (generally around -50 dBm.)

  • 802.11n is usually over 70% efficient. That is, your TCP/IPv4 throughput in a tool like iperf will be over 70% of your average PHY rate, and if you can hold the 300Mbps PHY rate (as is typical in perfect conditions) you should see well over 200Mbps. Of course a wireless-to-wireless transfer cuts that in half, so maybe that was your 90Mbps case? – Spiff Mar 12 '17 at 4:06
  • @Twisty Channel 1 at 40 MHz bonds 20MHz channels 1 and 5. Channel 11 at 40 MHz bonds 20MHz channels 11 and 7. So it has as much overlap as 20MHz channels on 5 and 7: 10MHz of overlap. So a quarter of your two 40MHz channels would be overlapping. – Spiff Mar 12 '17 at 4:12
  • @D'oh...of course. Channels 5 (and 7) would be the extension channels. I was trying to make it out that one would use a 40 MHz wide channel centered on channel 1 and 11, which is not correct. Need sleep... – Twisty Impersonator Mar 12 '17 at 4:21
  • @Spiff - you seem to have wandered right past the 50% loss in actual speed from it being only half-duplex. – Ecnerwal Mar 12 '17 at 15:28
  • @Ecnerwal No, the half duplex is why it's only 70+% efficient instead of 94% efficient like full duplex Ethernet. Network flows are not symmetric. Tiny 802.11 acks and TCP acks take a very small percentage of the airtime compared to the 1500 byte data frames of a download. – Spiff Mar 12 '17 at 17:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.