You are here

Mikrotik support

7 posts / 0 new
Last post
VA2XJM
Mikrotik support
Hi AREDN team,

I've discussed this morning with a possible provider about the project of an AREDN mesh using Mikrotik devices. From previous posts on the forum, Ubiquiti support might be dropped because of proprietary tech in their AC line, so I decided to look into Mikrotik.

Mikrotik also release their AC lineup of devices. Actually I feel that AREDN does seems to stand with the "N" standard devices. However the provider doubt the "N" devices will be around for a long time as availability is already a problem. He really encourage me to switch to AC devices for availability and also the fact that they are compatible with N network for existing nets. To me it feels normal to buy the latest tech so it can live longer before it needs to be upgraded.

I am also aware that Mikrotik is less supported than Ubiquiti so far by the project, but it may change in the future. So actually I have an eye for the following devices that I would be happy to order and help testing for official support:
  • GrooveA 52 ac
  • mANTBox 15s
  • LHG XL 5 ac
  • LDF 5 ac (Testing for very long link, no AREDN support needed for those tests)
But before investing the money into the required devices and equipments, it would be interesting to see the project roadmap and what exactly is planned in the future so wise decisions can be made by network admins. I guess it would also benefits to everyone to know exactly where the project is going with the new devices coming on the market.

I don't see any point fighting with my provider to special order 10x LHG XL HP5 (that I may not get them at all) when the AC version is ready to ship right now...

VA2XJM Jean-Michel

PS: It is not meant to be rude or disrespectful, I am sorry if you feels it is, my English is a bit limited...
AJ6GZ
More
These are more expensive but I think are also worth considering: RB921UAGS-5SHPacT-NM - NetMetal 5SHP or B922UAGS-5HPacD-NM - NetMetal 5 RB911G-5HPacD-QRT - QRT 5 ac (might be around longer) or RB911G-5HPnD-QRT - QRT 5 ...something made of metal, and something with a decent mount :) I would be happy to test both. Ian
K3MMB
Interested in point to multipoint
Would be glad to play

regards
mike
AE6XE
AE6XE's picture
Jean-Michel,   here is the
Jean-Michel,   here is the technical limitation we currently have...

802.11n chipsets use the linux wireless "ath9k" driver with a known way to change into ham radio channels
802.11ac chipsets use the linux wireless "ath10k" driver.   This driver is reported to not yet be 'hacked' to set channels outside of unlicensed part 15 channels:

ref:  https://wireless.wiki.kernel.org/en/users/drivers/ath10k

Known bugs/limitations:
"applying ath9k regulatory domain hack patch from OpenWRT causes firmware crash (reason: regulatory hint function is never called and ath10k never sends scan channel list to the firmware which in turn causes firmware to crash on scan)"

We can run on 802.11ac devices, but not yet in ham radio channels.  No one has investigated yet to see how easy/hard it is to get over this hurdle.   But I hope to spend some time on this in the next few months.

This is an issue impacting any vendor's hardware.  Note that Ubiquiti continues to manufacture and support the AirMax line, with new hardware revs coming out, e.g NanoStation M2 XW and Bullet M2 XW models recently  appearing.    There are some 802.11ac Ubiquiti models using the qualcom atheroes chipsets and compatible with the ath10k driver that are supported by openwrt and candidates for AREDN.   

Longer distance links are not able to get additional advantage from moving to 802.11ac.  This is because the new features in .ac are around more antennas and increased channel bandwidth.    This isn't useful in long distance links where SNR is a significant issue.   increasing bandwidth stretches the power across more space and lowers the SNR of individual carriers (64 in an 802.11 20MHz signal).   More antennas means more interference between spatial or polarized signals and overall doesn't work as well at long distances. 

BTW, there is an Mikrotik LDF 5 (802.11n) device about to arrive at my doorstep.   Hope to add this to the list shortly.

Joe AE6XE
k1ky
k1ky's picture
802.11 ac

Joe,

Thank you for the descriptions regarding long-haul aspects of 802.11 ac.  I had wondered how that was going to play out.  We usually have to drop back to 10 Mhz bandwidth for solid 5GHz links over 30 miles. Hopefully the new Receiver technology will help us drag signals out from under a lower noise floor?  Do you have any idea what the bottom end of these new devices will be compared to the -95 figure that we are dealing with currently?

AE6XE
AE6XE's picture
Noise Floor in 802.11
Noise Floor in 802.11

These devices don't have the physical ability to measure an absolute signal strength or absolute noise floor.   What they do to determine the noise floor is receive when there are no detected signals (Clear Channel Assessment -- CCA -- says no signals).  During this quite period they measure the noise level.   Then simply say this is -95dBm in 20MHz channel width.     Then received signals are assigned a level relative to this value.     Consequently,  if a given site had more ambient noise than another site, a known absolute signal level might be measured several dBm lower than another site.    

This default value is determined based on the hardware produced noise.   The math I saw was based on CMOS technology, so unless a different technology is used, receiver sensitivity stated values wouldn't change.    

The math to calculate this hardware noise figure includes factors for frequency, bandwidth, temperature.    When we cut the bandwidth in half, this should then be half the noise or -95dBm  - 3dBm =  -98dBm to use for the measured noise floor.  But the values shown with different channel widths don't account for this -- it's still not an absolute value, so someone probably decided not to bother.  

Where 802.11ac would benefit is the ability to send 2+ data streams out 2+ antennas to 2+ different clients at the same time.    The data stream on the vertical polarization could be going to one client and the horizontal polarization data stream to another client. 

with 2 antennas the V and H polarizations are 90deg apart.  with 3 antennas they would be 60deg apart and increased spatial interference.  at 4 antennas, they'd be 45deg apart, and so on.   We see better thoughput in 10Mhz channels than 20Mhz channels at longer distances, using 40Mhz or even 80Mhz would be increasingly worse thoughput at same distance.   Adding more polarizations would be similar decreasing thoughput (probably why it is hard to find higher gain antennas with more than 2 polarizations). 

Joe AE6XE
VA2XJM
Thank you for that clear and
Thank you for that clear and to the point reply Joe!

I will check back with providers to see if someone is wishing to sell gear or not... But so far the "N" devices seems to be hard to get in Canada as no one has them in stock and special orders are expected to be over 4weeks before availability.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer