Perhaps it would be better to call it "interference" vs "noise". But yes, you can have poor QoS due to interference, and still see the radio stats you're reporting. For example SNR is measured between received signal strength and average background noise level -- but if interference is bursty (which it typically is) it will not factor in the average background noise level. Similarly, interference will have no impact on the RSSI. This is because when the signal is received, it is always received at the same level. Problem is : sometimes it is not received! (due to interference drowning it out).PSA , when the DNS ping time goes from 3ms to 300ms , the signal strength , CCQ or signal to noise , non of them changes ! so it just goes from 3ms to 300ms all of the sudden without any change to any parameters !
and distance is 2km
Perhaps it would be better to call it "interference" vs "noise". But yes, you can have poor QoS due to interference, and still see the radio stats you're reporting. For example SNR is measured between received signal strength and average background noise level -- but if interference is bursty (which it typically is) it will not factor in the average background noise level. Similarly, interference will have no impact on the RSSI. This is because when the signal is received, it is always received at the same level. Problem is : sometimes it is not received! (due to interference drowning it out).PSA , when the DNS ping time goes from 3ms to 300ms , the signal strength , CCQ or signal to noise , non of them changes ! so it just goes from 3ms to 300ms all of the sudden without any change to any parameters !
and distance is 2km
Do a test:
Start an FTP download from an internet server (preferably a big file). Make sure you download it though the IP address of the server.
For example: ftp://48.25.26.145/pub/........ (fill in the IP address of the correct FTP server).
Run it when everything is "fine", when the DNS queries are 3 mSec.
note the speed, average.
Then run the same test when the DNS ping rises to 300. Do you still have the same speed?
Also, change the DNS servers to Google DNS: 8.8.8.8 and 8.8.4.4 or OpenDNS: 208.67.222.222
is he right ? is this a normal thing ?Since you are in an open frequency you can't count on you getting data across the links every single time so the ping time increases as the retry occurs.
Correct. The IP for the FTP server is an example.
this is a real one:
ftp://194.109.21.27/pub/test. There you will find a dummy file of 100 Mb or 1 Gb.
And yes, the DNS can be primary and secondary.
is this guy right ? this problem is a normal thing with wireless internet and has nothing to do with ISP ?!Since you are in an open frequency you can't count on you getting data across the links every single time so the ping time increases as the retry occurs.
Yes, he "can" be right. To investigate it will require a lot of effort, regarding finding out the "interferer", and even then, if you found it, what are you going to do about the guy (or woman) who messes up your link.....
Solutions:
1. Switch to 5 GHz. both ends should support it (the LHG is pretty cheap, 59$ per unit (you need two of them).
2. Switch to a different channel. Both sides should do that.
3. Use a more directive antenna, cover it with a dome if needed. But that might not work well, since the interference can also be at the ISP's end.
sorry if im asking too much i'm just trying to learn , but in your example , i can hear the noise right ? like in the street i can know that there is a noise coming from some where ( maybe not knowing the source , but at least knowing its because of noise)interference as an example: is when you talk to someone on the street, and there is some noise form cars, busses etc... When there is not too much, it is fine. When tere is a lot of car traffic, you should "shout" harder. And if you are in a disco or club, you really need to shout very very hard, and move your mouth close to the ear of your partner.
Well, you and your partner are the dishes of Mikrotik. The noise are other people.
Noise is always present, and you cannot avoid it. You can though make your signal so loud that the noise influence is minimized.
So OK, you are already on 5 GHz. Try to change the channel number. Not the bandwidth yet. The narrower the bandwidth the better for interference, but less throughput.
So do not change the channel width yet, but try different channels. 5 Ghz band has much more channels, and you will probably find a good one .
Let your isp to solve the problem for you. He should have the necessary experience and you can learn from him.
You can hear the noise. See the spectral scan from both ends and select the best frequency. Also use the dish antennas with side shielding to suppress the noise from the environment.
Well, I agree with Jarda.
Let the ISP solve it. If they cannot, then they are not capable in doing this. Select a different one (probably they are the only one with "high speed" internet), so then you are stuck with them...
You should isolate the problem first. Are you sure that the delay is on your last mile? Maybe it is somewhere in or behind the network of your isp. It's always better for consumer to buy the service delivered to his ethernet port leaving all the technology on the isp. Then its not your business what is wrong with the connectivity.... But its too late now. There is option to nicely negotiate with the isp to help you or to hire a local independent consultant to either help you to connect or to proove the origin of the problem. You cannot succeed with blind help over the discussion forum with this especially if you don't have own experience.
sorry for the late replyOk. There can be whatever. I told you to see spectral scan and to use side shielded antenna. May happen that it is badly aligned or there is noise on Ap side that makes your signal bad. Hard to say without local observation.
yes i checked google first ofcourse ! i saw that "the dude" program tooNot sure how it is on your place, but for me the google works fine:
https://www.google.cz/search?q=mikrotik ... 8wfnt4HgBA
first link:
http://wiki.mikrotik.com/wiki/Manual:Spectral_scan
Also check the spectral scan in the Dude...
thanks for the replySomething Jarda mentioned is very appropriate and hasn't been mentioned by either of you since:
The bottleneck might not be your wireless link at all.
Consider that whatever tower you're talking to, the ISP must have a good connection reaching that tower from the rest of their network. Your link could be the best one in the history of humanity with 0 packets lost in ten trillion packets, 160db signal to noise ratio, etc.... but if the ISP has problems reaching the tower you're connected to, then it won't matter.
Analogy: you could have a race track quality driveway and road in front of your house - it could be easy to drive 400km/hr on this with the right vehicle, and never crash, and there are never cars in the way, but if the road that your driveway connects to is busy with cars and has lots of potholes, then it doesn't matter how fast you can drive on your driveway, because the next road in your route to ANYWHERE will be slow....
Here's how to test this:
Use the traceroute tool in your SXT to see a list of IP addresses that your packets must travel through to reach some destination. Choose some well known address like 8.8.8.8 as your test destination. Then note the IP addresses in the first, second, and third hops.
In both good and bad conditions, try ping tests to each of these 3 addresses. If the first hop (most likely it will be at the tower you're connecting to) stays good even when pings to DNS server, and other ISP hops get bad, then your wireless is fine, and the problem exists further upstream.
Other clues to look for if the first hop's ping never gets worse:
second or third hop IP changes consistently with the performance changes (usually means that a link is failing to a backup link and getting congested)
First hop past your ISP's network starts getting bad pings (means the ISP's upstream bandwidth is congested)
My suggestion was basically what Paternot was telling you to do, but spelled out in steps... the basic idea was to determine whether your first hop is in fact the problem, or if it exists somewhere further up the line. It appears that your first hop is indeed the problem, but pings/traceroutes can only give clues, not definitive proof that the bandwidth is saturated. (i.e. running at 100% capacity)also thanks to Paternot for replying , but what do you mean by the bandwidth is saturated ? can you explain please and also do you think this actually could be our problem ? how to determine that this is the problem or not ?
I mean that they (the ISP) may have sold (too much) more bandwidth than they have. I'm thinking that maybe your signal is 100% okay, but there are too many users on the tower. In order to keep http "clean", they made a huge QOS queue. The result is exactly what You are complaining about: high latency and zero packet loss.also thanks to Paternot for replying , but what do you mean by the bandwidth is saturated ? can you explain please and also do you think this actually could be our problem ? how to determine that this is the problem or not ?
well my first hop is 192.168.1.1 and its the address of my radio and its fine with no problemMy suggestion was basically what Paternot was telling you to do, but spelled out in steps... the basic idea was to determine whether your first hop is in fact the problem, or if it exists somewhere further up the line. It appears that your first hop is indeed the problem, but pings/traceroutes can only give clues, not definitive proof that the bandwidth is saturated. (i.e. running at 100% capacity)also thanks to Paternot for replying , but what do you mean by the bandwidth is saturated ? can you explain please and also do you think this actually could be our problem ? how to determine that this is the problem or not ?
One thing to do if you can log into your own Mikrotik is watch the wireless connection statistics when things are bad - Is the CCQ dropping (it should be above 90%)? Is the speed dropping to low values? If these numbers start looking bad, then it means there's an issue with the wireless transmission itself. If not, then it means that the problem is either at the tower or somewhere further upstream from the antenna you're connected to.
As to how you don't see the next hop, this can be obscured in many ways, and it could even be the case that your connection goes through several devices and links before it reaches the next IP hop, and ping/traceroute cannot diagnose any of that infrastructure... However, if your first hop (above your own router of course) then that's something to complain about to the ISP.
yes the TX/RX stays stable aswellThe modulation rate is the speed that the radios are actually using to transmit to each other. I think they're called TX/RX rate in Mikrotik.
If these stay stable during poor performance times then the most likely cause for your problem is that the ISP's wireless sector is too oversubscribed- i.e. Has run out of bandwidth.
Judging from the spectrum scan, there's not much room in the area for another channel, either.
They are not necessarily lying.so you think my isp is lying to me ? because they are saying that it's because of the "Noise" from the area around my house !
so if they asked me how i'm sure that its not the noise , i should just tell them because that CCQ , Signal and TX/RX stays stable and in good condition during this problem right ?
also what are your thoughts about them using this IEEE 802.11a which is from 1999 ? could this be the problem ? that they are using old and outdated equipment ?