1

Topic: Whether can WiFi be reliable?

Now I use WiFi n. The modem is connected to an optical fiber from the provider (any GPON Lucent). On a receiving party  2012. Communication so-so: speed limit on  around 5 megabytes per second, not on  above 4 rarely rises. The provider gives 120 megabits and if by cable to the modem to be connected, basically it approximately so much and squeezes out. But the most terrible is unstable operation. All works as times normally,  to the modem within 3-5 msec, losses are not present, but losses of 10-20 % from time to time start to go, some packets go 50-100 msec (it at all I do not understand, for 50 msec it is possible to  to reach not that 5 meters on apartment) and for me it is critical since I play an online toy and at this time it it becomes simple . While I want to drag on apartment a cable, but for the certain reasons it to do difficult. If I distribute WiFi instead of the modem any more modern point of access that there was a standard ac, whether it is possible to expect, what it solves problems with a speed and ? The ether basically is not so littered, now, for example, it is visible only 14 networks. Everything that is necessary to me, it is fair gigabit on  and the guaranteed absence of losses of packets.

2

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: vsb> Everything that is necessary to me, it is fair gigabit on  and the guaranteed absence of losses of packets. Pull a cable.

3

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: Technology Wi-Fi (and generally a radio channel) cannot guarantee reliability by design. You want reliability - a cable. You want convenience - Wi-Fi. It generally. And in particular, many factors can operate. The ether ("an amount of networks" - not an index) can be littered; the signal can /// because of apartment singularities; the access point can switch in g a mode because of any incompatible devices (not mandatory yours), trying to it to be connected; there can be bugs in insertions; the piece of iron can is banal not to pull at the big traffic (game + ); etc. If it is ready to pay for convenience -  the expert, it measures all, finds the reason and makes recommendations.

4

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: vsb> Everything that is necessary to me, it is fair gigabit on  and the guaranteed absence of losses of packets. On 5GHz there are less than networks, than on 2.4GHz. If the router and a computer support this mode losses can disappear. I was helped by changeover of a router on supporting 5GHz.

5

Re: Whether can WiFi be reliable?

Hello, lpd, you wrote: lpd> On 5GHz there are less than networks, than on 2.4GHz. If the router and a computer support this mode losses can disappear. I was helped by changeover of a router on supporting 5GHz. Any more very much a variant... And such networks already becomes in bulk...

6

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: vsb> Now I use WiFi n. The modem is connected to an optical fiber from the provider (any GPON Lucent). On a receiving party  2012. Communication so-so: speed limit on  around 5 megabytes per second, not on  above 4 rarely rises. 5 MB/s == 40 Mbit/s It only on reception? And sends how many? It is necessary to look layout of the modem and , whether antennas  and walls are closed. At me, for example, the router in a guard costs - it was necessary to fasten to a board of the antenna and to drag in a router. Still it is necessary to look, in what mode really works WiFi (whether really it n?). As, speed depends  antennas in the modem and  and on an operation mode n (20MHz or 40MHz). I think, in  a router it is possible  to find this.

7

Re: Whether can WiFi be reliable?

Hello, andrey.desman, you wrote: AD> 5 MB/s == 40 Mbit/s AD> It only on reception? And sends how many? Approximately as much. AD> it is necessary to look layout of the modem and , whether antennas  and walls are closed. The modem costs in , from three sides of a wall. And so especially it is closed by nothing, in apartment of doors it is few. Any special difference between notebook layout it is noted. In  now writes Noise:-90 dBm, Tx Rate: 39 Mbps if it about something speaks. AD> still it is necessary to look, in what mode really works WiFi (whether really it n?). In properties of connection it is written 802.11n.

8

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: AD>> 5 MB/s == 40 Mbit/s AD>> It only on reception? And sends how many? vsb> Approximately as much. That is the general speed ~80/ (WiFi half duplex). + it is necessary to multiply on coefficient (I do not know what) since speed WiFi is underlined the general, instead of the data. Well, let us assume, on 90-100 Mbit\s it works. A maximum for n on one antenna - 150 Mbit\s. The question with an amount of antennas all as is actual. If to replace wifi the unit and a router on normal AC at least all will be much better. But warranties will be given only by a cable. AD>> it is necessary to look layout of the modem and , whether antennas  and walls are closed. vsb> the modem costs in , from three sides of a wall. And so especially it is closed by nothing, in apartment of doors it is few. Any special difference between notebook layout it is noted. In  now writes Noise:-90 dBm, Tx Rate: 39 Mbps if it about something speaks. Not so especially. AD>> Still it is necessary to look, in what mode really works WiFi (whether really it n?). vsb> In properties of connection it is written 802.11n. So it is better in  than a router to look, there it is more .

9

Re: Whether can WiFi be reliable?

vsb> Everything that is necessary to me, it is fair gigabit on  and the guaranteed absence of losses of packets. Nobody gives a warranty to you (an ether - a piece gentle). And to receive fair gigabit, it is necessary to work and mighty well to be enclosed in the equipment. Start to read from here, well and the reality (unlike the theory under the link above) looks [ur = https://www.smallnetbuilder.com/old-too … rofile-dn] approximately so [/url], that is with one router and one network card currently it is better not to calculate more than on 500 mbit/with, and on the average and at all on 300-350.

10

Re: Whether can WiFi be reliable?

Hello, Nikolay_Ch, you wrote: lpd>> On 5GHz there are less than networks, than on 2.4GHz. If the router and a computer support this mode losses can disappear. I was helped by changeover of a router on supporting 5GHz. N_C> Any more very much a variant... And such networks already becomes in bulk... 5GHz has much more smaller penetrating ability so from other networks strongly it is less than noises.

11

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: vsb> But the most terrible is unstable operation. All works as times normally,  to the modem within 3-5 msec, losses are not present, but losses of 10-20 % from time to time start to go, some packets go 50-100 msec (it at all I do not understand, for 50 msec it is possible to  to reach not that on apartment) vsb> If I will distribute 5 meters WiFi instead of the modem any more modern point of access that there was a standard ac, whether it is possible to expect, what it solves problems with a speed and ? Most likely yes. "Rollback" on 802.11g by loss of speed Still can help a little stability raises. vsb> the ether basically is not so littered, now, for example, it is visible only 14 networks. 14 networks in 5 ac? For 2.4 14 networks it already is a lot of, especially if at them levels of signals it is better-60. So that at me in 2.4 of a window is visible more than 120 points of access, but at the majority of them the signal is worse-70. What in dynamics shows the ether scanner, for example https://lizardsystems.com/wi-fi-scanner/? To compare time when communication normal and time when communication bad.

12

Re: Whether can WiFi be reliable?

Hello, vsb, you wrote: vsb> Now I use WiFi, some packets go 50-100 msec There is such problem, studied it in details as wanted too most - game with control on vaj-faj through a router vaj-faj. A trouble with a time delay. Problem researches showed, very detailed research... That the modern insertions of routers are guilty. The time delay is imported by forcedly program method. On loading of an ether depends very little. The time delay is and when at me on the channel of 30 networks of different capacity (19 hours) and when I one (4 o'clock in the morning) Clearing up of the reasons was led by testing different vaj-faj routers in the same conditions. Outputs - all NEW i.e. approximately since 2014 vaj-faj routers import a time delay, and even purposely lose packets if to hollow them  50 times in 1 second Vgljadit as attempt to buffer frequent  an ether for the purpose of preventing of swamping of the channel by clients. Or something something like that. As you know since 7  it is possible to make  vaj-faj a router on vaj-faj the PC. Since 7 SP 1 such points also specially import a time delay. Though it is considerable smaller than vaj-faj routers. And the algorithm is more difficult - adaptive, on empty channels the time delay is less and it  skips. In routers  I  - a time delay fixed,  and different depending on an insertion and model. Old routers admit  to 200 times a second lost-free even at rather loaded channel. Here this old vaj-faj the router steadily gives very small time delay and is suitable for GAMES and control we tell the joystick on vaj-faj. The time delay is small even at 20 networks on the channel from neighbour's routers. D-Link DVA-G3672B if find such - safely take.  for games without . Search of adjustments responsible for a time delay did not lead to success. Did not find them. Even for  7 though it is obvious they is, as  I am much visible simply could not combine  parameters of routers so to disconnect a time delay. OUTPUTS: MODERN vaj-faj routers do not suit games with short delay Old D-Links - suit.

13

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: L> the Time delay is and when at me on the channel of 30 networks of different capacity (19 hours) and when I one (4 o'clock in the morning) At 4 o'clock in the morning they are ungeared physically? They generally are not present on air or you think that time people sleeps, networks do not occupy frequency? And different independent devices  do not swing,  do not wake up, updates do not swing? L> clearing up of the reasons was led by testing different vaj-faj routers in the same conditions. L> outputs - all NEW i.e. approximately since 2014 vaj-faj routers import a time delay, and even purposely L> lose packets if to hollow them  50 times in 1 second That such to "hollow"? ?  and any icmp the traffic has the least priority, the processor for answers to it is given on residual to a principle. It generally can not answer on , and it is normal. L>  for games without . Simply switch the new router in a mode "only 802.11g". L> OUTPUTS: Experiment is delivered not correctly, the theoretical basis is not brought, outputs are erratic. New routers by default work in 802.11n. In this mode, for capacity growth (those "300 ") it is used MIMO aggregation of channels, simultaneously goes transmission and reception on two frequency channels (still there is MIMO to magnification of an amount of antennas, but it for another). Accordingly and noises from "neighbors" are caught not on one channel in a band of 20 Mgts, and in two, in a band 40 (but, see a picture more low). Each badly accepted packet . As, in new wifi chips sensitivity is refined. If earlier, to old chips the signal-65 actually was not audible, the new hear and  to hold more or less a stable relation even at-80. From this follows that they are forced to process and noises with level-85, not only on the but also on adjacent channels. Working on channels 1+6, to process noises (and the useful signal) on 1,2,5, 6,7 channels. Taking into account more broad band and the best sensitivity, in the conditions of a typical city (when each coffee maker creates ), noises it is caught essentially more is more often (than in 802.11g), arrive beaten packets, , the general ACTUAL throughput is more often is necessary more low. Switched in 802.11g you you narrow down working frequency and you lift a threshold of cutting off of noises. That not to make including 802.11n on operation in one frequency channel, there there is such possibility for struggle against an amount of noises, but all the same it is necessary to process low-level noises of adjacent channels (5, 6,7 for example).

14

Re: Whether can WiFi be reliable?

SK> That such to "hollow"? All is easier and more difficult... Old vaj-faj routers, I was specific specified D-Link model ALWAYS work without time delays. In the conditions of noises simply part of packets does not reach (UDP) the more strongly a noise the more often do not reach. Thus time delays skip from a minimum to 100 msec and are arranged on the Gauss where peak proportional to force of noises (obviously game tries to be punched by repeated sendings) New ALWAYS work with a time delay. Without dependence from level of noises. This cardinally different behavior. It was tested nearby 20 vaj-faj routers. Correlation with insertion versions is revealed 100 % - it should be till 2014 That's all. All remaining theories of simply theory. Practice (which criterion of the validity) speaks - the small time delay for games (UDP) is necessary take the model specified by me or older (D-Link DVA 7236B) and  to itself brains any far-fetched confirmed REAL  theories. My hypothesis - many insertions are based possibly on any open-sorse, possibly author in 2014 imported there algorithm of a time delay and after that it spread on many insertions of different devices. Therefore the question - under Windows 7-10   an access point becomes, the time delay there is. With  it is necessary to understand through  , I did not master, but it is a lot of adjustments and it is probable to eat there time delay regulation. The question to unixoids - whether whether is on your favourite axis a software for the organization of a point of access from  and if there is that there is in it a time delay on UDP at klient1-> (vaj-faj a router on )-> 2

15

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: SK>> That such to "hollow"? L> all is easier and more difficult... Old vaj-faj routers, I was specific specified D-Link model ALWAYS work without time delays. I attentively read all opus, but and did not see the made answer to a question "That such" to hollow "?" L> correlation with insertion versions Is revealed 100 % - it should be till 2014 L> That's all. In 2014 officially accepted the standard 802.11n, before it was in draft copies. Hence, till 2014 in wifi access points it is by default switched on 802.11g instead of 802.11n it was necessary to switch manually. With 2014 on the contrary. L> and  to itself brains any far-fetched confirmed REAL  theories. Comes loginx to restaurant, sees on desktop a dash he the waiter asks what is it? it answers  members measured, well loginx gets puts, so at me the biggest! The waiter: at first the muzhik remove a term and secondly they measured on the other hand!! L> the question to unixoids - whether is on your favourite axis a software for the organization of a point of access from  and if there is a software for the organization of a point of access from a desktop does not approach? Whether L> that is available in it a time delay on UDP at klient1-> (vaj-faj a router on )-> 2 it is absolutely senseless question while you adequately will not answer a question "That such" to hollow "?"

16

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: L> It was tested nearby 20 vaj-faj routers. A technique you will not share? How measured, than generated the traffic etc. to Repeat probably?

17

Re: Whether can WiFi be reliable?

Hello, wildwind, you wrote: W> Hello, loginx, you wrote: L>> It was tested nearby 20 vaj-faj routers. W> a technique you will not share? How measured, than generated the traffic etc. to Repeat probably? Yes, you take 2  (or the PC with vaj-faj), on on one client UDP on other the server,  to vaj-faj. You press the button  (changes color),  as quickly color changes a receiver of relative change of color , for pseudo measurement I removed both screens (they nearby) on the camera of 240 frames/sek More precisely  to me and it was not necessary as time delays were huge, actually and without shootings by eyes it is perfectly visible. P.S. The problem was found out when wanted to play simple Ping-Pong with  1 on  2 through vaj-faj a router  it very much , it would be desirable to find nevertheless those parameters that it is necessary to twist in the modern routers that worked on a time delay as old D-Links

18

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: W>> the Technique you will not share? How measured, than generated the traffic etc. to Repeat probably? L> yes, you take 2 PCs, on on one client UDP on other the server,  to vaj-faj. L> you Press the button  (changes color),  as quickly color changes a receiver of relative change of color , L> for pseudo measurement I removed both screens (they nearby) on the camera of 240 frames/sek L> More precisely  to me and it was not necessary as time delays were huge, actually and without shootings by eyes it is perfectly visible. One udp-package? What for the client and the server? Source codes are?

19

Re: Whether can WiFi be reliable?

Hello, andrey.desman, you wrote: AD> Hello, loginx, you wrote: W>>> the Technique you will not share? How measured, than generated the traffic etc. to Repeat probably? L>> yes, you take 2 PCs, on on one client UDP on other the server,  to vaj-faj. L>> you Press the button  (changes color),  as quickly color changes a receiver of relative change of color , L>> for pseudo measurement I removed both screens (they nearby) on the camera of 240 frames/sek L>> More precisely  to me and it was not necessary as time delays were huge, actually and without shootings by eyes it is perfectly visible. AD> One udp-package? What for the client and the server? Source codes are? Source codes are everywhere in Google, I took ready of Google for C# both on Java and on Delfi - the pattern is invariable changeover of a router by an old D-link instantly solves a problem. While it satisfied me, as is in a warehouse a little old vaj-faj routers and need while is closed. But if who understands more deeply but not the theory and in practice and solves a problem  or adjustments  is cool. And yes, violent  buffers, the big or small buffers - all it does not change a pattern cardinally. P.S. Certainly   with a router wires also at once does a time delay hardly noticeable in my circuit of measurements.

20

Re: Whether can WiFi be reliable?

Hello, Stanislaw K, you wrote: whether L>> that is available in it a time delay on UDP at klient1-> (vaj-faj a router on )-> 2 SK> it is absolutely senseless question while you adequately will not answer a question "That such" to hollow "?" To send UDP frequency we tell 200 times in second Well I do not wait for anything intelligible from the person not guessed that such to hollow. You had any other versions about to hollow, except how often to send? Sound simply for the sake of interest that you there to yourselves imagined?

21

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: whether L>>> that is available in it a time delay on UDP at klient1-> (vaj-faj a router on )-> 2 SK>> it is absolutely senseless question while you adequately will not answer a question "That such" to hollow "?" L> to send UDP frequency we tell 200 times in second Well I do not wait for anything intelligible from the person not guessed that such to hollow. L> you had any other versions about to hollow, except how often to send? I on former still had question "That such" to hollow "?". I will set in addition leading questions: Whence to send? Where to send? What size? Whether it is exact UDP? Whether to wait for the answer? Specially for senior pupils a question in pictures:" You hollow "from a notebook where? We assume on  there can be what that the answer (but after all it at all the fact, and so, the maiden dreams). Variants: the smart phone should Answer? What there at the smart phone with productivity? Has time to accept, process, answer? Adjustments firewall? Other notebook on wifi should answer? What there at a notebook with productivity? Has time to accept, process, answer? Adjustments firewall? Other computer on a wire should answer? What there at other computer with productivity? Has time to accept, process, answer? Adjustments firewall? The router should answer? What at a router with productivity? The router should answer with LAN the interface? The router should answer with wan the interface? What there at a router with adjustments firewall? At suspicion on dos attack  packets? At what excess pps? Who should answer that on the Internet? What at it ACTUALLY with productivity? What at it with firewall? Answers from it are returned by what route? (It after all IP, packets there and packets can reversely walk very different routes, tcp certainly adds a few order, but you after all insist on UDP) L> Sound simply for the sake of interest that you there to yourselves imagined? What for to me free of charge to dream. I at all did not begin, as you and did not answer "That such" to hollow "?".

22

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: W>> the Technique you will not share? How measured, than generated the traffic etc. to Repeat probably? L> yes, you take 2  (or the PC with vaj-faj), on on one client UDP on other the server,  to vaj-faj. L> you Press the button  (changes color),  as quickly color changes a receiver of relative change of color , and, tested new routers do not consult to send a packet of 200 times in a second?

23

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: L> yes, you take 2  (or the PC with vaj-faj), on on one client UDP on other the server,  to vaj-faj. L> you Press the button  (changes color),  as quickly color changes a receiver of relative change of color , L> for pseudo measurement I removed both screens (they nearby) on the camera of 240 frames/sek L> More precisely  to me and it was not necessary as time delays were huge, actually and without shootings by eyes it is perfectly visible. Interesting. And what was a time delay? And what dispersion on models, from "new"? P.S. Fast  showed that appears there are routers which are presented as "game".

24

Re: Whether can WiFi be reliable?

Hello, loginx, you wrote: SK>> That such to "hollow"? L> all is easier and more difficult... Old vaj-faj routers, I was specific specified D-Link model ALWAYS work without time delays. As there was: http://lurkmore.to/%D0%9D%D0%B8_%D0%B5% … 0%B2%D0%B0 In routers there are no artificial time delays. Most likely, it is simple  all , but the old router used inaccessible (under the law) in Russia channels. L> and  to itself brains any far-fetched confirmed REAL  theories. Ah. REAL experience, a pancake. I easy play Overwatch with a time delay less 1 from my local WiFi (the full time delay - 20) and a stable relation. Differences with Ethernet are not present. ?

25

Re: Whether can WiFi be reliable?

Hello, wildwind, you wrote: L>> yes, you take 2  (or the PC with vaj-faj), on on one client UDP on other the server,  to vaj-faj. L>> you Press the button  (changes color),  as quickly color changes a receiver of relative change of color , L>> for pseudo measurement I removed both screens (they nearby) on the camera of 240 frames/sek L>> More precisely  to me and it was not necessary as time delays were huge, actually and without shootings by eyes it is perfectly visible. W> it is interesting. And what was a time delay? And what dispersion on models, from "new"? W> P.S. Fast  showed that appears there are routers which are presented as "game". At game 20, at type super fast super modern -  buffer type, accumulates 5-10-20-30-50 packets then spits out and if packets are similar or identical that can only last (or a little), the algorithm is similar on  i.e. It is impossible to state that something is mandatory lost, but also instantly does not reach... Yes, the method  packets on a path klient1->> a router-> an ether-> 2 for not game routers does an ether - their few-convenient in games. I will repeat for inattentive - the effect takes place at  klient1-> an ether-> a router-> an ether-> 2 in the presence of wires all works quickly and without  effects. The remark of the previous orator is probable matters - the most game router of D-Link DVA-G3672B has only 802.11b/g