Expired Hot Deals

[Best Buy] TP-Link HS200 Wi-Fi Smart Light Switch $24.99

  • Last Updated:
  • Dec 23rd, 2018 2:22 pm
Sr. Member
User avatar
Jan 7, 2014
866 posts
245 upvotes
calgary
AntonioG23218 wrote:
Dec 20th, 2018 1:29 pm
Makes sense, thanks. Now I feel better.
thats why you need to monitor RFD 24 hrs a day so you dont miss any good deals
Deal Addict
Mar 18, 2006
1648 posts
529 upvotes
BC
blue dragon wrote:
Dec 20th, 2018 9:14 am

Now you are mixing up CS/MACD from ethernet with RF transmission mechanism from wireless. 12 devices transmitted in the order of 300kb/day (150kb down and 150kb up), The amount of time that these devices are actually transmitting is minuscule.
Your response makes no sense and you'll need to express a response for me to understand what you're trying to say. WiFi is listen before talk protocol. Two devices don't hear each other, then transmit at same time and interfere. This is because of propagation delay, one transmits before receiving the other signal to know to not transmit.

300Kb per day. You think the clients hold hands and know when the guy next to him is talking, waits his turn and only ever talks when no one else does?

You talk about having wireless captures all day long, but you don't see the constant packets doing probe requests even already being associated to the AP. There's nonstop chatter, all day long. That makes me think you're not wirelessly capturing and only seeing filtered data packets.
Koodo: $40/5GB Unlimited Canada wide calling
Deal Addict
User avatar
Sep 13, 2006
3271 posts
2752 upvotes
Muskoka
With respect to lost connection issues

I had a couple of devices that would sometimes lose connection and I would have to reset or power cycle. For other reasons I upgraded my WiFi system to the TPLink Deco 5 mesh system. As a side benefit all of my devices became quicker response and I don’t think have lost connection since. My Logitech Harmony Hub is also more responsive.

So I am sure the quality of your router has something to with it
Koodo
Public Mobile
Member
Jun 21, 2009
208 posts
171 upvotes
Timbo420 wrote:
Dec 21st, 2018 5:45 pm
Your response makes no sense and you'll need to express a response for me to understand what you're trying to say. WiFi is listen before talk protocol. Two devices don't hear each other, then transmit at same time and interfere. This is because of propagation delay, one transmits before receiving the other signal to know to not transmit.
What I am saying is the time take to transmit a few bytes is miniscule. Even though there was contention, the actual channel utilization is small enough that the contention mechanism will be able to find free time for other devices to transmit. As for the probes, the probe request sent by an already authenticated endpoint every few seconds are on different channels, not the same channel that it is connected with. Are you mixing up beaconing from the AP which transmits the SSID and BSSID and occurs on the same channel with the probes sent by clients on different channels?
300Kb per day. You think the clients hold hands and know when the guy next to him is talking, waits his turn and only ever talks when no one else does?

You talk about having wireless captures all day long, but you don't see the constant packets doing probe requests even already being associated to the AP. There's nonstop chatter, all day long. That makes me think you're not wirelessly capturing and only seeing filtered data packets.
I never said I am doing wireless captures, I said I have a packet capture for data leaving my network. My Meraki APs do give me the ability to do a wireless capture, but the only time I did a wireless capture was when I used the air magnet tool to verify my network after provisioning. No one is saying there aren't probes every few seconds, but you imply they will cause your wifi network to crawl, and that simply isn't true. Those probes are so that endpoints can discover other APs with better signal strength, other SSIDs etc.

As I said before, wireless is a contention medium, that does not mean 12 devices will not break it. 802.11 DCF implemented random backoff timers for endpoints transmitting at the same time.

Image

https://www.cisco.com/c/en/us/td/docs/s ... n_ch2.html

I can show you APs with 20 devices connected simultaneously delivering RTP (UDP) traffic for Skype for Business. Yes, QoS is configured, but I have QoS configured at home as well too.

Those TP-Link switches all use TCP which has re-transmission capability, windowing, and acklowledgement mechanisms to deal with a noisy or congested link.

Is your position that 12 devices is enough to bring your wifi network to a crawl, or to significantly impair its performance as the OP stated? Even Zigbee is a contention network. Only RF networks like TMDA or FDMA etc solve the contentiion issue
Deal Addict
Mar 18, 2006
1648 posts
529 upvotes
BC
blue dragon wrote:
Dec 22nd, 2018 4:14 pm
What I am saying is the time take to transmit a few bytes is miniscule. Even though there was contention, the actual channel utilization is small enough that the contention mechanism will be able to find free time for other devices to transmit. As for the probes, the probe request sent by an already authenticated endpoint every few seconds are on different channels, not the same channel that it is connected with. Are you mixing up beaconing from the AP which transmits the SSID and BSSID and occurs on the same channel with the probes sent by clients on different channels?


I never said I am doing wireless captures, I said I have a packet capture for data leaving my network. My Meraki APs do give me the ability to do a wireless capture, but the only time I did a wireless capture was when I used the air magnet tool to verify my network after provisioning. No one is saying there aren't probes every few seconds, but you imply they will cause your wifi network to crawl, and that simply isn't true. Those probes are so that endpoints can discover other APs with better signal strength, other SSIDs etc.

As I said before, wireless is a contention medium, that does not mean 12 devices will not break it. 802.11 DCF implemented random backoff timers for endpoints transmitting at the same time.

Image

https://www.cisco.com/c/en/us/td/docs/s ... n_ch2.html

I can show you APs with 20 devices connected simultaneously delivering RTP (UDP) traffic for Skype for Business. Yes, QoS is configured, but I have QoS configured at home as well too.

Those TP-Link switches all use TCP which has re-transmission capability, windowing, and acklowledgement mechanisms to deal with a noisy or congested link.

Is your position that 12 devices is enough to bring your wifi network to a crawl, or to significantly impair its performance as the OP stated? Even Zigbee is a contention network. Only RF networks like TMDA or FDMA etc solve the contentiion issue
No. My position is that you were claiming interference is only when your network is significantly impaired, and I was saying there is always interference and that is expected and designed into the protocol. The amount of interference one can live with varies with application. Not everything runs on TCP, and even when it does, it hugely impacted by delays and dropped packets. That's why some APs tried to prioritize TCP acks in their buffers, so TCP streams wouldn't see large drops in speed.

I think we both agree it's minimal to negligible impact, but disagree on what constitutes "interference". It varies but will always be there.

Lastly, how a client operates for roaming and finding other AP's is OS specific (and power saving modes), you'll find it sends probes on same channel for different ssid's (it wouldn't have a reason to probe for its currently associated SSID on same channel). At least Android does and so does BB10. But I'm actually thinking of mobile devices with shit loads of SSID'S and not single ssid IoT devices, so that's my bad, it's less relevant to this IoT as additional devices discussion.
Koodo: $40/5GB Unlimited Canada wide calling

Top