Yep, that's already happening with almost any kind of APs available these days. However in this case a brief disconnect happens when client roams from one AP to another, which is a no-go for some applications like VoIP.Do you mean as you move from one AP to another if they have the same SSID + Security (which CAPsMAN manages) then it should auto switch to the new AP?
Because this already happens.
And is usually more-so governed by the client device rather than the AP.
If you mean another type of roaming, please explain it more clearly.
I understand the request and it is a good one, but just wanted to note, that you can already configure access list to disconnect client with bad signal, and the client will then reconnect to the nearest AP
FALSEFirst of all - roaming is always a client decision; When to roam, what AP to roam to, what band (2.4 or 5) to prefer.
FALSEnb: I say badly but the reality is that MT is designed as a fixed location PTP/PTMP endpoint so should never roam
And that is where the confusion begins. By kicking off the client (a deauth request), the AP is simply telling the client to disconnect from it. The client may decide to reconnect to the same AP or try another AP, based on it's association algorithms.FALSEFirst of all - roaming is always a client decision; When to roam, what AP to roam to, what band (2.4 or 5) to prefer.
AP can kick a client if weak signal, max clients reached or member of load balancing group forcing roam.
Good for you, it just means your client devices are very good roamers.FALSEnb: I say badly but the reality is that MT is designed as a fixed location PTP/PTMP endpoint so should never roam
I have 44 radios (88 SSID's) connected to a single CCR1036 (capsman+usemanager) and i can have a perfect skype conversation between all SSID's roaming seamless in 2.4 and 5GHz.
If it works well for your application, go for it. The point of this discussion is different. OP asked that "CAPsManager should support Roaming between networks" and a lot of misinformation ensued (including yours I'm afraid), so I wanted to set the record straight on how 802.11 WiFi works.I liked the term "reactive roaming" you have said before. I understand your concept and I'm sure there are better formulas to roam but mikrotik works very well and for a lot less money than any other solution .
Not really. Roaming is the way 802.11 is designed to work!. In fact, Apple devices tend to be the best behaved devices on a WiFi network and their roaming behaviour is well documented - http://support.apple.com/en-us/HT203068This prevent Apple device to constantly jump, register, disconnect over detected APs when in idle mode. This completely kills the bandwidth in a multi-AP environment...
This is exactly what is it about.Pretty sure station roaming allows the MikroTik to roam when it is in station mode.... e.g. it does nothing if the mikrotik is acting as an access point.
If you actually understand the request: why do you sugest using the braindead (drop the client in the cold) ACL approach. A real wifi controller verifies that another AP is actually hearing the client and reception is better, then it will try use appropriate protocols for moving client if it supports it otherwise disconnect but then you know that the client is actually in range of another AP instead of leaving it out in the cold with no service at all instead of a low connection.I understand the request and it is a good one, but just wanted to note, that you can already configure access list to disconnect client with bad signal, and the client will then reconnect to the nearest AP
No, this is station roaming, for MikroTik stations only (not phones, laptops, etc.)Obviously someting was added in the latest 6.39 rc4 release...
I got all excited there for a secondNo, this is station roaming, for MikroTik stations only (not phones, laptops, etc.)Obviously someting was added in the latest 6.39 rc4 release...
Ubiquiti's "Zero Handoff" feature is different from what people here are asking for. It is effectively a "single channel" system like Meru/wiNG. Even Ubiquiti do not recommend it!Just wanted to say for everyone asking for roaming implementation like Ubiquiti has, did anyone actually tried it so far? I have several of this UniFI APs at one customer, and i tried zero hand off feature and made everything so much worse and unstable that i had to turn it off, it simple doesn't work as good as you would expect it..
Well yea its brand new thing introduced this year, i was just saying about zero handoff which was originally requested prior to 2017, im sure mikrotik will give us something similar to FastRoaming soon(i hope)Ubiquiti's "Zero Handoff" feature is different from what people here are asking for. It is effectively a "single channel" system like Meru/wiNG. Even Ubiquiti do not recommend it!Just wanted to say for everyone asking for roaming implementation like Ubiquiti has, did anyone actually tried it so far? I have several of this UniFI APs at one customer, and i tried zero hand off feature and made everything so much worse and unstable that i had to turn it off, it simple doesn't work as good as you would expect it..
What people here are asking for is an equivalent to Ubiquiti's Fast-Roaming https://help.ubnt.com/hc/en-us/articles/115004662107
Those standards are relatively recent/new and the number of client devices that are actually capable of using those is greatly exaggerated. Also they still do not allow the AP to dictate the client when and where to roam- that still is totally client's decision. The standards you are referring to just (potentially) make the roaming process faster/more efficient. But please note that clients that support (and can benefit from the use of) these standards are likely to be rather good roamers even on their own (i.e. without AP assistance).Some people argued here that this is client responsibility to swith AP. Yes, it is. But client logic relies on AP supported standards, like 802.11r/k.
Try reducing the tx-power of your APs? That helps, really helps.Without ACL workaround clients dont switch APs when they should.
Sounds reasonable. I will definitely try. As I suppose AP Tx power should correlate with clients Tx power.Try reducing the tx-power of your APs? That helps, really helps.
In my opinion only client should decide when it should roam. Not AP. Thats why I do not like ACL workaround, as I described in my first post.Also they still do not allow the AP to dictate the client when and where to roam- that still is totally client's decision.
Look around. What does people really want? Whey want to use WhatsApp/Viber/Skype/etc voice/video calls. They use iphone/ipad/android with latest firmware. Vast majority of these devices do support these standards (https://support.apple.com/en-gb/HT202628)Those standards are relatively recent/new and the number of client devices that are actually capable of using those is greatly exaggerated.
I'm getting rather good results with 50mW output power, which means setting tx-power to 15dBm for 802.11ac-capable interfaces (where tx-power means total power) and 12dBm for double-chain (10dBm for triple-chain) non-802.11ac-capable interfaces (where tx-power means power per chain). You may also consider decreasing the tx-power for 2.4GHz interfaces even further to give 5GHz band some priority.Would you recommended Tx power for me (I am a home user: iphone, android, laptops etc) for channels 1, 6, 11 and for 5G?
We also wanted to participate in this project to extend our infrastructure. It seems, EU money will go to another company. Perhaps Mikrotik don't need this money?The project requirements for WiFi4EU are:
(..)
support IEEE 802.11r
(..)
But unfortunately Microtik does not meet the requirements.
For Germany, Italy, Romania,France, Spain, the funds allocated for project WiFi4EU is 20 160 000 EURO. Mikrotik lost that money because they dont need this money. Another company example: ub*** ,cis***** etc. cover all requirements. MIkrotik why not wake up ?We also wanted to participate in this project to extend our infrastructure. It seems, EU money will go to another company. Perhaps Mikrotik don't need this money?The project requirements for WiFi4EU are:
(..)
support IEEE 802.11r
(..)
But unfortunately Microtik does not meet the requirements.
Mikrotik is on Wireless one way road with old legacy kernel and own wireless driver with mostly basic hardware support....For Germany, Italy, Romania,France, Spain, the funds allocated for project WiFi4EU is 20 160 000 EURO. Mikrotik lost that money because they dont need this money. Another company example: ub*** ,cis***** etc. cover all requirements. MIkrotik why not wake up ?We also wanted to participate in this project to extend our infrastructure. It seems, EU money will go to another company. Perhaps Mikrotik don't need this money?The project requirements for WiFi4EU are:
(..)
support IEEE 802.11r
(..)
But unfortunately Microtik does not meet the requirements.
Great news! In which version of ROS6? Is that only announcement of the development or some builds are available as of now?Since lately there is some developement in ROS6 features regarding long awaited features, I would like to bump this one.
- dot11k/r/v based roaming / AP steering in CAPsManager
- dot11k/r/v based band steering (also without CAPsManager)
- option for airtime fairness
Also from other threads:
- support of wave2 (MU-MIMO, beamforming); viewtopic.php?t=145047
- tunable DTIM and beacon interval; viewtopic.php?t=116221
- stop broadcasting device model and ROS version with the beacon; viewtopic.php?t=133186
- support dot11d including country code; viewtopic.php?t=72840; viewtopic.php?t=144151
Great news! In which version of ROS6? Is that only announcement of the development or some builds are available as of now?
Oh I see now. A hastily & hopeful interpretation of that by me, to the point that I upgraded multiple devices to v6.45.b62!!. So still waiting and hoping....at least some more attention might be drawn here reincarnating this wish list....Don't get me wrong! I was hopeful when dot1x and ike2 features were released in 6.45beta, that developement of often reqested features was sped up. So I brought up my favorites to give them some attention.