How WiFi Works - Computerphile
Vložit
- čas přidán 29. 05. 2024
- Why do WiFI speeds vary so much? Dr Steve Bagley demonstrates how even a poor signal for one person can affect those with a seemingly perfect signal!
/ computerphile
/ computer_phile
This video was filmed and edited by Sean Riley.
Computer Science at the University of Nottingham: bit.ly/nottscomputer
Computerphile is a sister project to Brady Haran's Numberphile. More at www.bradyharan.com
Me: Yeah I know how wi-fi works
Steve: Explains how wi-fi really works
Me: Ok then, I'm glad I watched it
😄
just open example .conf file for "hostapd" next time you think you know it.
@@Mr.Leeroy Yeah no, I already know what's in it.
Wow, I've never thought about the fact that Wi-Fi routers can only transmit one packet at a time to a single receiver. And the final example with two people using the same network and the person with a slow connection results in a slow connection for everyone -- it was just mindblowing!
Bear in mind, this issue is solved in modern devices through a method called OFDMA, which allows for full-duplex (both ways simultaneously) communication between multiple connected devices and the host.
@@thelegalsystem Eh, I think you mean MIMO. Using every antenna to speak to individual clients.
The best thing is to disable PSK beacons (1, 2, 5.5 and 11 Mbps) and use OFDM (802.11a/g) speeds
@@tkteun yeah my bad, I get my consumer wireless and carrier wireless terms crossed up a bit.
@@thelegalsystem I don't think it's Full Duplex. But WiFi6 has MU-MIMO(on top of OFDM) this allows simultaneous down link packets to go to multiple users at the same time. Requires multiple antennas to steer beams towards users and(if implemented) nulls towards interference.
@@scienteer3562 Beamforming is a whole topic in itself. And I thought the multi user was already in 802.11AC / Wifi5, while 802.11N / Wifi4 was the first with MIMO (ie. no longer a dedicated send and receive antenna)
guys stop posting interesting topics when i am currently watching something else which i am also interested in
Yeah, been a network engineer for a decade, studied and worked with multiple wireless data transmission technologies, and whenever anyone asks how wifi works, my stock response is "surprisingly well, when it's not working poorly". What they really want to know is "how can I fix my wifi speed", and most of the time my suggestion is to switch off the 2.4GHz radio in your router and just use the 5GHz one, because there's a lot more free channels so a better chance it will pick one that's not congested. And then if the reception is poor in the furthest reaches of the house, it's really time to invest in a better wifi solution.
But at least now if they really want to know how wifi works, I can just point them to this video and save myself the 20 minute explanation!
At home we’ve switched over to 5GHz, much better speeds than 2.4GHz since we live in an apartment building. And if there’s any device that doesn’t support 5GHz then I use Ethernet (like for some of my MANY Raspberry Pi:s).
but then came wifi 6E with 6GHz and started using wide bands which effectively reduces the number of available and separate bands.
I find it best to think about Energy required to receive a single bit. You can increase the transmitted Energy per Bit by either slowing it down, or increasing the Power.
Surely slowing down would decrease energy?
@@sandworm9528 if you half the data rate, each bit is transmitted for twice as long. Energy = Power x Time .
@@scienteer3562 gotcha, cheers
This is a great explainer! Do OFDM and MIMO next :)
THIS. THIS . THIS
And MU-MIMO
Yeeeeeees
Don't forget QAM and CSMA/CA
This was fascinating. I had no idea about this aspect of wifi. I feel like this could use a better title than the generic "how wifi works". I almost didn't click on it because I thought it'd be another basics I already knew. This is more like "how your bad signal can slow everyone's wifi".
One thing I didn't think of for a long while. There's usually only one radio receiver in a laptop/phone etc which can only listen to one channel at a time. Yet when you click the WiFi button it has to show you all the different base stations, many of which are on different channels. So your device retunes the radio for about a tenth of a second, per second through each of the different channels to find all the base stations. But while it's doing that, it misses packets on the channel it's actually connected to, so they have to be resent! That's why the announcement is every tenth of a second. So there's a bit of jitteriness on WiFi that you don't get with Ethernet.
Added complication: those channels overlap. For example, with the original 802.11b on 2.4GHz, there are supposed to be 11 channels, but in fact there is only room for 3 entirely separate channels -- 1, 6 and 11. So all the others overlap one of these (and each other) to quite a large extent.
@@lawrencedoliveiro9104 1,5,9,(13) don't overlap significantly though in the common cases where they're not precisely colocated.
That is intresting. I guess that also explains why network discovering can take so many seconds and why they are found in such random feeling cadance.
@@lawrencedoliveiro9104 Wi-Fi is seriously complicated.
Atari ST in the background? Thats some legit cred right there :)
💯
My favourite way of explaining these concepts is using food dye in water. The more dye, the more power. The further out from the origin point of the dye in the water, the further from the wifi router. Add lots of different food dye colors in represents interference from other devices, which you can overcome by using more power (putting more dye in), but eventually you're so saturated that it would take unreasonable amounts of dye to get something representing color, instead of a gross brown/black.
~30 years in IT. Still learned a lot! Thanks for this video!
this should have 10 million subs by now - love computerphile.
Thanks for uploading this now, my networks exam is tomorrow! Good luck Imperial Computing second year, hope this video finds you
Great video as always. Please talk about DHT some time!
THANKS for providing all these videos!!!!
Very interesting video
I'm now playing with the idea that all of my slow IoT objects in my house (Google Home Minis, Ikea outlets, etc.) should say use the 2.4Ghz network while my laptops and phone should be "isolated" on the 5Ghz network as the far away IoT objects might be slowing down the network?
Not sure what an ikea outlet is but I assume that (like their smart lightbulbs) those actually operate on bluetooth, not wifi, connecting to a bridge device which is plugged in with an ethernet cable. In this case, I don't think those particular devices should affect your internet speeds much.
If they’re sharing the same band, then they’re contributing to interference and reducing your bandwidth, even if they’re using an entirely separate protocol.
it's possible, but keep in mind your iot devices are probably using very little bandwidth. Even if their signal is poor the amount of time they spend actually transmitting
eceiving is low.
the other factor is 2.4ghz generally has much better range and penetrating power than 5ghz.
2.4ghz also tends to be much more congested (less people run 5ghz routers and devices).
as you can tell it's .....complicated.. you could test it for yourself, but the results could still be subject to change (due to the congestion factor changing, based on what your neighbors do or don't.
They should also really be on their own VLAN for security
@@TAP7a Assuming you aren't using only locally communicating devices, yes. Personally, it made no sense to me to communicate to an Internet server to communicate a signal from one device on my LAN to another device on my LAN. I went the Tasmota and LocalTuya route (along with Home Assistant, which is...fine); highly recommended for those willing to make the effort.
Thank you for explaining this. I had a few different experimental SSIDs being broadcast by my AP, each tied to a different VLAN. Now that I know that it's more-or-less radio pollution (because I'm not actively using those SSIDs), I've turned them off until I'm ready to resume experimentation.
love you guys please never stop making videos
Fascinating video, lovely stuff as ever 😁
Love to see a video on wifi channels if it's not already done
I actually do think how this stuff works and knowing this channel background, it's gonna be a pretty deep dive , thanks
Dr Steve is a national treasure
Thank you for that video.
I see a trend to use WiFi for everything. Don't use cable, it is something of the past, use WiFi, it is so convenient, event for the desktop of the secretary, that is sitting on their desk and will never move anywhere, even for that huge printer that obviously needs a cable for main power.
You demonstrate clearly that using WiFi when it is not needed is a bad idea (i.e. slow)
Saturation is biggest issue with most wireless networks
I prefer cable. It's more reliable, consistent and faster.
Tell that to my IPTV decoder I got from my ISP. No Wifi... they don't TRUST wifi to be strong enough. Only cablese.
But you can use ethernet over powerlines!
10:27 - A small correction if you'll allow me to be a pedant for just a moment. I think the presenter meant to use the term Beacon frame rather than Beacon packet when talking about network discovery. The distinction is minor, but frames and packets are different, have different header/trailer information, refer to a different layer of the OSI model, and have other implications.
So if the access point is sending its SSID beacon 10 times a second at its slowest speed, does this noticeably impact the overall connection speeds? Does disabling SSID broadcasting (which I think is possible on most routers) potentially improve performance for devices that are already connected?
I was wondering the same thing. Also was wondering if the standards group(s) should reduce how often beacon's are broadcast. How often does that really need to be sent? Yikes.
@@JeremyMcMahan Yeah, the 10 times per second really surprised me. But someone pointed out that your laptop or phone can only tune into one frequency and thus will have to scan all channels by listening to each one for 100-ish milliseconds.
If 15 personal hotspots slow down wifi to a crawl, the beacon takes up one fifteenth of the bandwidth, would be my guess.
Nice video learned something new. Can you explain more about ZigBee vs Bluetooth, Mesh network that uses a 3rd 5ghz band and upcoming wifi 7.
The last third of this video is really informative
Gonna need a half hour special with the puppy as special guest.
This is satisfying explanation
Can you do one on WiFi Mesh systems like Google wifi?
Great advice! :D
So would making a hotspot SSID hidden be faster as it would but have to broadcast itself every 102.4mS at the slow speed & free up that bandwidth?
In my case I would not care that it was hidden as I always connect to my own network which is already known to me.
I really feel like this was the point of the video, and they just didn't address it.
3:46 “These are SMALL. But the ones out there are FAR AWAY...”
smaaLLLL...
....Farrrrr Awaayyyyyyyyyyy
The second part about two people on the same network was explained well, but the first part about the beacon packets from hotspots slowing down neighbouring networks was quite quickly glossed over. Did I miss any part of this video, or did the gentleman not explain *why* the addiitional beacon packets affect other networks?
I didn't get the part as well. I think it was implied that everything gets flooded with "I am a router, this is the info to connect to me" messages and the router for some reason has to take turns with the hotspots and wait, or something like that.
Basically all wireless devices would listen medium before putting packets on air. They will have predetermined threshold values for signal and energy. In case of more packets on medium this signal detected on medium would be high which may crosses the threshold value at sender. Hence, it would defer for sending packet out which results in slow speed.
For example, say we have 2 host (wireless device) on network when host 1 wants to send data, it will only account the signal from host 2 and it would defer based on that. But, if i added host 3 to the same band then host 1 will account signal from both host 2 + host 3 which would definitely higer than former case.
if you are sending bigger packets, that means the longer time is being spent sending those packets which in turn means the longer time the router IS NOT able to send packets to others.
@@Ddub1083 Yes. This correct in case of Single user MIMO. With Multi user MIMO router can serve multiple client at the same time.
@@manikandanramaraj3434 sure but it does so at the cost of splitting up the bandwidth... so you arent really gaining anything unless you are just splitting between several low bandwidth uses. instead of it being slow some of the times with congestion, its half the speed all the time but consistent.
7:35 I thought I was losing my mind... **woof**
How is the speed negotiated? I'm assuming similarly to the beacon packet, at a known speed, so that it can be clocked? Or possibly a known preamble so the clock can be calculated from that?
3:37 "these cows are small. But the cows out there... are far away" .
argh, that lovely ST at the back
Great video but if you're going to do it remotely PLEASE improve the audio!
audio was fine ?
Isn’t there such a thing as an echo-cancellation microphone? The front one picks up your voice, the back one picks up the room reverb, and by subtracting the latter from the former, you get a much clearer, less boomy recording.
Any resource to see/read anything and everything about Wi-Fi's and how they work?
Xfinity/Comcast in the past few years implemented broadcasting its own guest xfinitywifi ssid in addition to the home or business network on their modems and don't have a way to turn off the xfinitywifi unless you physically replace the modem with your own (however I haven't tested for that ssid in bridge mode so there may be a way to do that if you have your own wifi router behind it). Considering apartment buildings or condos where a lot of those can be in close proximity it is just aggravating that the rental modems that everyone seems to get are cluttering the airspace.
Nice Father Ted reference there! :D
So you are saying you could load 50 ESP8266 ICs on a pcb, put them all in APSTA mode, connect them all together and it could render an areas wifi usless?
This a great video, very well explained and I learned a lot. But sound quality is really bad, maybe a new mic would do the job. Thanks!
If you’re using a hidden SSID is it still broadcasting an beacon packet?
Hidden wifi won't broadcast beacons and so they don't show up in passive scans, they will only respond to "probe request" attributed to their SSID.
There will still be a beacon frame. It won't contain the SSID tho.
Hiding the SSID has no benefits. It forces all WiFi clients that want to connect to go like "Hey (SSID here) can u hear me? Im searching for you" all the time when they are not connected. Even when you are absolutely not in range.
If your network is "static" in terms of no mobile devices that disconnect from time to time, you can still get the SSID by deauthenticating one of the clients and forcing it to reconnect that way. When connecting, it will send the SSID.
@@sporqist thx
What@@sporqist said! One end or the other has to initiate the broadcast so the other can find it. Better for it to be the access point than your mobile device. Otherwise your device will be broadcasting the name of your home network wherever you take it.
There was this peculiar thing called the “Free Public Wi-Fi” bug, back in the days of Windows XP, as I recall; all these devices were going around trying to connect to a network with that name, and nobody could figure out why. As one device broadcast that name, others would pick it up and try to connect to it too, and so the “infection” would spread.
All I need is a laptop,a certain Wi-Fi antenna, and a few Linux commands to drop network activity to nothing? Good to know.
Signal jamming can be done even easier
Why you gotta post videos during finals week ;)
How well secure are dem open sources Linux distros...... dr Steve?
I love how you recreated the 1990s production quality feel!
I remember going outside of the house and after almost 100 metres I noticed I was still connected only to wifi and I could still watch youtube
crazy shit, it normally didn't do that
do wireless access point continuously send out beacons ,until some client establishes connection?
Without beacon packets, your device wouldn't be able to find the wifi network. Even after you connect to the network, those beacon packets still need to be sent out. If those beacon packets stop, then it won't be possible for you or anyone else to connect to the network, as it will not be found.
What affect do channels have?
Can you make a video on why to use CRT in 2022 ?
just WOW
How does it get fasted or slower while still using the same frequency? Does it repeat or something?
They take longer sending each bit. Same carrier frequency, but longer bits on the carrier.
When area is saturated with WIFI signals - isn't proper channel selection designed to aleviate this problem? When setting up a WIFI - access point listens in promiscuitie mode for foreign wifi networks and should automatically select the best channel based on the 802.11 b / g /n specification and lowest utilization, ideally a channel that is not used at all?
As discussed in the video, if one is increasing the time interval of the packets, to account the distance from the router. Will the tome interval automatically decrease when the user physically moves closer to the router?
Also, how is the time interval icreased per packet anyway?
Kindly explain.
Mostly the packets will have same time. only the length would vary. For the short payload, sender would do padding. For example a standard OFDM symbol would take 4us (3.6us data + 0.4 us guard interval) on 802.11a case.
To answer your question, Here increasing and decreasing the time means How much time taken by AccessPoint to send the data to client. When the Client is closer to AP the chances of getting errorred or corrupted packet is low. So, AP will put more data bits per symbol (collective of subcarriers) hence speed increases. Where as if the client is far away from AP then chances of getting errorred or corrupted packet is high. So, AP will put less data into the symbol. Hence, it will take more time to transmit whatever data it has for client.
Here, the error could be error vector magnitude which is nothing the difference between actual and error point on constellation map at the receiver side.
Sender would always monitor the SNR ( Noise floor - received SSI). If it became less from the expected then it will reduce the modulation. Please refer MCS with SNR requirement for each standard.
l want some of the Good Doctor's "Things". They sound like fun
When the wifi has been idle for a while it will go down, right? And only one of the clients would be able to get it going again. I'd had a server connected to a local network by wifi, but the system broke after a while. Cable between the access point and the server solved the problem.
Interesting to think hotspot tethering could be used for DoS
You say that an access point can only talk to one client at a time, but can't it use a band of frequencies and allot one to each client, thus reducing interference this way?
I don't think so because the router can either transmit or receive with one client at a time. MIMO technology overcomes this by having multiple transmitters and receivers. Thus, not only can it communicate with more than one client at a time, it can also transmit and receive simultaneously with any given device.
So question to anyone who knows: will making the wifi network hidden or turning off the beacon open up speed on the network? (Pardon my lack of knowledge if the question doesn't make sense)
Do ipv6 full video.
Nice
one single slow wifi connection will slow down all other traffic on the same access point ? I didn't know it !
How would beacon packets sent by a hotspot interfere with those sent by a public router? Wouldn't the public router just keep sending its packets once every xxx ms (say) regardless? Why would it slow down?
Because on any particular frequency, only one of the devices can transmit at a time. Each one goes every 104ms, or just under ten times per second. Each additional AP is doing the same. They stagger their beacons, so now you have two or three or more every 104ms. It doesn't take too many before the time beacons are being transmitted becomes most of the time available for any traffic.
@@strehlow Ah I see - thanks Scott!
This sounds somehow exploitable to me: is there a way to send slow packets to stop the WiFi functioning, just like Slow Loris Attack? Is there any way to protect ourselves facing this?
Wondered about the same and I don't think you can prevent it. It is illegal to do so but you can just send gibberish across all channels, effectively blocking them. Just see "Wi-Fi jammers", there are even tutorials on CZcams.
@@teasin951 From what I've seen, these Wi-Fi jammers use deauthentication like aireplay-ng? I don't know any that uses this kind of passive way of attack. Just like Slow Loris, if this is actually implemented, it would be harder to detect than plain old deauthentication packets.
So a good solution to have fast WIFI at home is to build a faradays cage around your home and ONLY relay signals from the outside via cables that you are actually interested in. Like phone signals for example.
Then you have a good network of hotspots that sit close to any wireless device so none of them goes into slow mode.
...
Come to think of it... It would probably make more sense to just have everything connected to a cable.
Which is the sole reason i don't use WIFI at home. There's hundreds of hotspots here and everyone suffers slow speeds on their WIFI.
hello marley!
If the signal gets wider (or slower) for longer distances to be able to send clear data, wouldn't the receiver just get bits that are duplicated?
Like for example if the access point is sending 101, when the signal becomes longer wouldnt it be interpreted as 1111000001111? Or is there some sort of mechanism to calculate the time and adjust them accordingly?
I would assume that the access point and client negotiate the 'speed' beforehand.
It's called symbol encoding. In RF communication, symbols are not just encoded as 0:low level 1: high level. They use more sophisticated modulations for synchronization and redundancy. Just google "WiFi encoding and modulations", you'll get a lot of materials to dig in. You'll find out WiFi actually uses different combinations of encoding and modulations under the hood for various circumstances.
I assume its similar to USB, there is 1, 0, J and K.
Now of course wifi is not differential, but Im sure there are tricks.
Like the header may contain a byte that if organized a certain way will go back to renegotiate.
Or if the error checking fails a certain amount of times it could also go back.
You will get clear picture by referring MCS(modulation and coding scheme) index with SNR requirement of various standards.
Nah, it reads 1s or 0s only when the signal changes its state.
Noice, 17 minutes well spent :)
I thought each user had his own frequency or something.. as each have his own FM channel?
That... dose look like a design fault..? I am no expert.
In any case, dose making my wifi invisible boost my performance?
Or limiting to only strong connection to connect, to keep, say all cars driving at 80 km/h? And no odd ball congestion-oining at 30 km/h?
Is this collision problem the reason for why i cant be in a wifi network and create a hotspot at the same time from my phone?
That sounds like an OS protocol stack limitation.
It may be due to a limitation on whatever wifi card your device uses.
@@wlockuz4467Are there cards that are made for connecting to and or creating multiple wifi networks? (Guess he already mentioned the beacons at the university). Also relaying data / access from one wifi into another wifi would be cool on a more mobile device.
keeping skype alive
Will there be a "how GPS works"
Please upgrade your microphone or tune its settings :)
using wifi to understand wifi 😎
So much learning from this channel.
I think I've a solid grasp on a topic
until I watch a video here and become
re-educated / updated / indoctrinated.
Hey, Computerphile. Do a video on why my printer is always out of ink
Wow
How does wi-fi work?
Well, it's radio waves, innit.
Right when I started to watch this video my WIFI got significantly worse. Weird
Australian scientists invented WiFi. Wi-Fi uses a large number of patents held by many different organizations. In April 2009, 14 technology companies agreed to pay Australia's CSIRO $1 billion for infringements on CSIRO patents. Australia claims Wi-Fi is an Australian invention, at the time the subject of a little controversy. CSIRO won a further $220 million settlement for Wi-Fi patent-infringements in 2012, with global firms in the United States required to pay CSIRO licensing rights estimated at an additional $1 billion in royalties.
Maybe thats why ESP is so cheap. They arent paying royalties.
I sometimes run into this trouble when the wifi tether suddenly slow down and disconnecting. Then checking with wifi analyzer, my wifi hotspot is using the same channel with several wifi hotspot >3-4 near me. I would've guess that the software would've adjusted if a channel have more than 2 hotspot. Lol nope. Now we know why.
What really pisses me off: I set the channel manually - and then all the routers in the neighbourhood who have the channel set to automatic move to it over time. I don't know why they do it (well, I kinda do, because the ISP is notorious for their cheap and bad rental devices - still doesn't explain what's going wrong on the firmware level). But it means I have to switch the channel every few weeks.
102.4 ms is a nice computer number 2^10
Am I the only one trying to work out which Atari ST that is in the background?
Atari STe :)
@@DrSteveBagley Thanks, I can sleep easy tonight! 😁 Knew it wasn't a Falcon, think you have one of those in your office at work.
👍
Can you do a video about free internet vpns
For security, you should ALWAYS turn SSID off. Not only does that speed up your network by a slight amount, it also means that your neighbors don't know you have a wireless access point. Have a guest over that needs access? Give them your network name and it can connect. It's sad how few people know this.
I actually had someone stealing my internet access in an apartment complex. Put an end to that right the heck away, i can tell you that! That was before i really looked into WiFi, which I still think is a stupid name for wireless access.
Security through obscurity
@hk I do both. My wireless network is about as secure as i know how to make it. Which, granted, won't stop a real hacker (i'm looking at YOU NSA), but it will slow them down a bit. At least i hope.
And it should keep script kiddies out for sure.
Hiding your SSID actually degrades the network. The AP do still transmit beacons, except the name is set to NULL. In addition, clients need to constantly send a probe request when you hide your SSID, in which the AP needs to respond, these contribute more to the network congestion. Security wise, it doesn't really provide much protection. For average users like your neighbor, not being able to see your SSID might provide marginal security, but it really makes no difference if you are using a secure password without hiding your SSID. For hackers, hiding your SSID is totally useless. Once a client associates with your AP, your SSID will be revealed. The bottom line is it's better to use a highly secure password instead.
Food for algorithm ;-)
Why does the wi-fi signal get crap with distance?
Inverse square law, innit.
Long story short: use wired networks! Thanks!
How wifi works-
me- it doesnt thats my problem!
then address the problem. ?
In the 2.4 GHz space having a rogue AP on an adjacent channel is almost an order of magnitude worse than a rogue AP in the same channel (the reason network engineers get distressed when they see anything other than channels 1, 5 and 11 used)
Thankfully 5GHz only has a tiny amount of channel overlap (yes there is but nobody likes it being mentioned.)
Almost the first thing that should be done on any Wi-Fi config is to disallow any data rate below 12MbS in part to have faster beacons, but in reality it's to prevent any 802.11b device from dragging things into the mire both speedwise and it's awful CTS issues
Other interesting things such as having hidden SSIDs on the 5GHz on the DFS channels (ones which have to coexist with weather radar.)
A certain huge semiconductor manufacturer decided to (badly) implement Listen Before Talk (LBT) so their devices would never connect to hidden SSIDs on the DFS channels because they couldn't see the SSID but would never break radio silence to send a probe on those channels.
^^ 1, 6 & 11
@@aaronpower8741 Alternatively you can go 1, 5, 9, 13, if you have 13 channels available. (And I just realised that I have it memorized incorrectly as 1, 5, *8,* 13. That's gonna take some time to get rid of.)
You said what
@@rolfs2165 Yes. I'm Australian, so we do have Ch 13 here. Although I think you will find that was only really useful on older B and G networks. With N onwards, they use pretty much the full channel width and you still get *some* interference between adjacent channels. Happy to be corrected on that if I'm wrong.
I heard your mic is wifi XD
I live in a log cabin with 2ft thick walls and the house has a metal roof. Needless to say WiFi (and cell phone signals) are 'problematic'...I might as well be living in a Faraday Cage...
Should two-foot thick logs make that much difference? Unless their moisture content is a bit high ...
@@lawrencedoliveiro9104 it's 200 years old and has concrete between the logs...
@@lawrencedoliveiro9104 There's enough water in Victorian-era limestone walls to seriously impede wifi signal - I imagine the logs are even wetter
Wireless transmission - esp WiFi - sounds like a mess.
So in conclusion dont game on wifi because you will cause everyone else gaming having to wait for your slow wifi.
Known since middle-ages but still ppl do it. Highly annoying.
I made video about WiFi, simple explanation.
Wifi works like ethernet.... just wireless. Only one communication at a time is sent over the medium (cable vs air). Its just that the differences in time to send to various computers in different environment is vastly more affected by air than on the cable.
Personally I woulda done it differently, even when wifi was 1st made ram was relatively large on average, I woulda just had a full block of memory that record every "pixel" of noise and compare every generated string using those blocks against valid request strings, if it's valid then process it as would normally (refuse if no session and setup/close session etc), then use that same "pixel" of noise to communicate back. To be clear here I mean the band of noise, so 1Hz vs 2Hz or whatever way the signal is represented in words, the point is that the devices will only support upto to a certain precision of noise analysis, that is what I'm referring to as a "pixel" and flat out recording maybe 32 blocks of it or something should be more than do-able with the amount of ram that could & probably is built into them, a few dozen pooled threads with dedicated string buffers & flags etc should be enough to keep the analysis fast even when the number of "pixels" is in the 1000s to 10s of 1000s, just wait for them to finish manipulating their strings with the new block with simple appendage & shifting etc to check for a valid request string (which should be sufficiently long/complex enough to rule out coincidence, 32 characters minimum should be fine I think), whatever the results they would simply be fobbed off to other pooled threads for each session/connection while the main thread loads/reads/copies the next set of "pixels" to fob off to the analysis threads.