I found plenty of recommendations in the internet to "buy MikroTik router and few UBNT APs". Why is that? Is it some super nice couple that works great or MT doesn't have equivalent product to ubnt? Or Unifi APs are somehow super awesome, better than ones MT offers?

I probably would try to keep single manufacturer in network, so why to mix UBNT with MT?

I found plenty of recommendations in the internet to "buy MikroTik router and few UBNT APs". Why is that? Is it some super nice couple that works great or MT doesn't have equivalent product to ubnt? Or Unifi APs are somehow super awesome, better than ones MT offers?

I probably would try to keep single manufacturer in network, so why to mix UBNT with MT?

Up till recently, it's correct that for a pure access-point Ubiquiti had some interesting offers Mikrotik could not directly match.

Recently Mikrotik has started investing into CAPsMAN (Central controller like UniFi) and for instance the hAP AC which basically uses about the same hardware as the Ubiquti UAP-AC-PRO and works well as a pure access-point. Mikrotik's new wAP AC will soon be available too, making a nice access-point for ceiling and wall mounting.

So in the past, yes, but recently they are very much catching up. I've been using hAP AC's for a bit now and they function very well in a CAPsMAN setup.

Oh that makes sense. Thanks.

are you talking about wifi wLAN or wisp wireless??

Indoor APs. Unifi precisely. I found plenty of recommendations.

But wasn't eg. RB951 competitive? I mean well there are and were plenty of decent wireless hardware offered by MT including beefy RB2011. So I'm not really sure why now hAP AC is competitive but back then in lets say 2012 there was nothing competitive. I don't really remember old MT offer because my first MT router was RB2011 so around year 2014' I guess? But it's almost hard to believe now that it used to be different just few years ago.

Which products are "equivalents" of Unifi APs in terms of indoors wifi quality (not appearance, signal stability/range preferred over speed)? I mean apart from hAP AC because it's actually kind of expensive, the same price as RB2011 which is used by me as primary router at the moment, I'm talking about rather lower end non-AC Unifis.

I bought hAP lite recently as second AP but unfortunately it doesn't seem to have really powerfull radio . Or at least it's completely smashed by RB. I mean currently RB2011 has stronger signal at any distance above like 2-3m from hAP so hAP provides rather negligible coverage. Considering crowded block of flats it's quite useless. Sometimes even devices in the same room choose RB to connect instead of hAP staying like 3m away. But still nice device for 20$. Well at this price tag it's almost overpowered

But I was in general interested why people suggested to use ubnt instead of MT APs for network with MT routers. So i guess it's been answered - historical reasons. I thought that maybe I'm missing something super "special" about ubnt.

I bought hAP lite recently as second AP but unfortunately it doesn't seem to have really powerfull radio . Or at least it's completely smashed by RB. I mean currently RB2011 has stronger signal at any distance above like 2-3m from hAP so hAP provides rather negligible coverage. Considering crowded block of flats it's quite useless.

the differences between tx power and rx sensitivity are well referenced on datasheets:
as we can see there is a difference of 8-10dbm on tx power and 1-9dmb of rx sensitivity

keep in mind 9dbm is 3x times exponential difference !!! :
20dbm=100mw
30dbm=1000mw

16dbm=40mw
25dbm=320mw

happened to me some time ago with the rb951-2n, it has even worst numbers, and learned the hard way as you

this way i never forget the impact and importance of this specs on coverage performance

off course for 25US hap lite is a very good performer

keep in mind most smartphones operate at 10-16dbm tx power then using and ap at very high tx power can help to improve downloading data-rates but the upload data-rate is determined by client device tx power and AP rx sensitivity.

to this info subtract the environment attenuation loss and interference and you get the real world performance

at least in mikrotik products you have this honest and detailed specs, unifi only publish the maximum tx power and in some cases this value is not accomplished in practice

keep in mind most smartphones operate at 10-16dbm tx power then using and ap at very high tx power can help to improve downloading data-rates but the upload data-rate is determined by client device tx power and AP rx sensitivity.

This is a very wrong assumption. For the best performance (both up- and download) you need that both your AP and your smartphone/laptop/tablet/whatever have comparable tx power. Higher tx power is of no use indoors, if you need better coverage the only way to go is adding additional APs.

keep in mind most smartphones operate at 10-16dbm tx power then using and ap at very high tx power can help to improve downloading data-rates but the upload data-rate is determined by client device tx power and AP rx sensitivity.

This is a very wrong assumption. For the best performance (both up- and download) you need that both your AP and your smartphone/laptop/tablet/whatever have comparable tx power. Higher tx power is of no use indoors, if you need better coverage the only way to go is adding additional APs.

you have reason, the better and good practice for indoor wifi is to use low power and more AP's to increase capacity and give good performance, but is a frequent practice by many many people to increase TX power on AP trying to solve some kind of problem.

In much cases increasing tx power can improve a little the situation because the download data-rate increases in some cases improving performance

off course if upload data-rate is as bad to disconnect the client or insufficient to return the ack's to the AP, increasing tx power on AP do not solve the problem

the other reason is cost. Mikrotik outdoor APs are more expensive than ubiquiti's but at the very least all mikrotik hardware run the same OS whereas ubiquiti's products sometimes have a software flaw or some flaw that doesnt make them great for some situations.

So i suggest either mikrotik/ubiquiti based on the situation and its never always a mikrotik router with ubiquiti AP.

What i do like about mikrotik's indoor APs like some of the RB9xx is that they have mini PCIe so you can upgrade them and do all sorts of stuff. You could make your own AC3200 but with 3 radios instead (including integrated) for less than half the price of whats on the market but it requires you to get all the parts together which many people just dont want to go through the effort or learn routerOS. Mikrotik however dont have a case for the RB9xx with 2 mini PCIe wifi cards as you would need antennas and support for mini PCIe cards are fairly limited from mikrotik. If you used a mini PCIe SSD i dont even think it would work. You would also need to do LACP on the wired side to keep up with the bandwidth of 3 radios but those things only have 2 ports, 1 of them being SFP. While you could still do LACP SFP adds a bit of cost. If there were 2 ethernet and 1 SFP it would help.

mini PCIe when used with an adapter can do great things.

There is the fact that despite CapsMan... we can't get competitive WAPs from Mikrotik into the US currently.

the other reason is cost. Mikrotik outdoor APs are more expensive than ubiquiti's but at the very least all mikrotik hardware run the same OS whereas ubiquiti's products sometimes have a software flaw or some flaw that doesnt make them great for some situations.

Is it? I mean, maybe I'm comparing wrong products but to me prices of outdoor are quite similar. But I don't really know much about outdoor so I can't really tell if UBNT is more affordable.

However indoor Unifi APs actually seem to be more expensive than MT (again unless I'm comparing wrong products)

you have reason, the better and good practice for indoor wifi is to use low power and more AP's to increase capacity and give good performance, but is a frequent practice by many many people to increase TX power on AP trying to solve some kind of problem.

In much cases increasing tx power can improve a little the situation because the download data-rate increases in some cases improving performance

off course if upload data-rate is as bad to disconnect the client or insufficient to return the ack's to the AP, increasing tx power on AP do not solve the problem

I experience the same with indoor APs. I've been replacing RB751Ui-2HnD APs with 2 or 3 hAP Lites and the results are much better especially in offices where there are a lot of walls/partitions. hAP Lites are very economical. Definitely getting much better user experience.

As for UBNT, some of my customers like the flying saucer UAP-AC design better. So that's why I go with that sometimes. Most of them don't like the cAP design so I use wAP indoors very often.

I only got 1 wAP to play with. Sitting right next to the 951g... it's throughput is right around 1/2. This is to the same clients. On the same channel. AFTER TURNING OFF ONE OR THE OTHER. Both are managed by capsman. Distance to test devices... 14 feet with no obstructions.

If wAP AC can equal or better the UAP AC PRO (not my favorite at all), and be sold in the US, I would have a place for them in my installs.

As far as APs go UBNT has more features than mikrotik such as zero handoff (their new AC ones dont though). But if wireless hardware features arent a concern and this is for indoor use i would choose the RB9xx series as they have miniPCIe so you can upgrade or make a monster of a config and beat the AC3200 or AC5300 by having 3 5 Ghz radios.

The good thing about mikrotik relying on software is that their APs can handle hundreds of clients whereas ubnt APs can have trouble handling more than 15.

My $0.02's worth is that Mikrotik lags behind in the dual band gear unless you want to build a dual-band box from a routerboard and a PCIe expansion and enclosure, antennas, pigtails, etc.

hAP ac is very competetive w/ unifi but in the U.S., the regulatory issues have mad things annoying for 5ghz Mikrotik gear.

I think Unifi is more evolved feature-wise but their server requires you to have an extra component in your installation.

I dont see why you would have issues with 5Ghz in the US, according to wikipedia all of the 5Ghz channels are available, probably power output. Mikrotik firmware is already locked so i dont see how there should be any issues with the fcc.

I was before among those supporters of the idea to combine Mikrotik Routers (RB2011 in my case) with UBNT APs. My first UniFi AP was non AC version (bought somewhere in 2012 or 2013 I believe) and it ran smoothly and cool for more than a year. Then I decided to invest in first gen AP AC, cause 2,4 GHz band was really overcrowded at my house. Mikrotik had nothing competitive at that time, so it was merely nothing to choose from. UniFi AP AC was 3 times more expensive in comparison to "normal" UniFi AP, but not worth the money at all! It is running very hot under load and produces buzzing sound when transmitting data. I'm really disappointed with this device and I'm thinking about replacement for it. Now UBNT have 3 devices in upgraded AP AC line, and even Lite version could be enough for my needs, but I would go for Mikrotik's wAP AC now (with upgrade of RB2011 to RB3011 in the future). Because of my previous experience with UniFi AP AC and because I like RouterOS very much.
And if you can run with all-in-one than you can try hap AC.

Hello,
I own the router Mikrotik CCR1009-8G-1S-1S + PC and WiFi sustème UBNT UAP-AC-Pro
Everything works fine without problem
Claude

Ubiquiti Unifi UAP AC PRO + RB750Gr3 is perfect combination for my house. Highly recomended !

Recently few quite exotic client infrastructures and issues had led me to conclusion that any network device that doesn't allow running scripts (without additional Linux box) is usually realistically useless... I was also really surprised by fact that ROS ability to run user scripts on events/schedule is actually not "standard feature" present in all Cisco/Junipier/UBNT/etc devices. It's funny really... So funny that almost miserable...

My $0.02 is simple:
Mikrotik is cheap, powerful, and infinitely flexible network equipment.
Mikrotik CAPsMAN is a beast to setup initially, and pushing configurations/using it is bleh.
Once you have a Unifi controller setup once, it's pretty, functional, powerful and so easy to manage all your AP's you could hand the admin panel to a user and they'd understand it.
Unifi takes all the work out of channels/radios/bonding/guest isolation and just works.

My 2 cents... For some reason, i found that UBNT devices do a better job at bridging. So that makes it for me, 2 ubnt devices for link, Mtik Routers at the endpoints.

My 2 cents... For some reason, i found that UBNT devices do a better job at bridging. So that makes it for me, 2 ubnt devices for link, Mtik Routers at the endpoints.

How? You mean performance wise? Recently ROS moved to hw accelerated bridging from switch master port. Also there's fastpath and fasttrack. Feature wise there's full L3 firewall on bridge level which is life saver for me. Is there some magical feature that you find really useful in UBNT?

I've chosen Unifi APs (first LR and then later AC Pro) primary for their looks when mounted on the wall. cAPs and wAPs are just ugly.
And sadly, MT has nothing to offer on the PoE switch side that comes near to ES and US (yet, I hope)...

I've chosen Unifi APs (first LR and then later AC Pro) primary for their looks when mounted on the wall. cAPs and wAPs are just ugly.
And sadly, MT has nothing to offer on the PoE switch side that comes near to ES and US (yet, I hope)...

Really?

when it comes to installations without poe switch i prefer MT products for the poe cable that is included. with ubnt you need small extra lan cables 0.5m to connect their psu.

I've chosen Unifi APs (first LR and then later AC Pro) primary for their looks when mounted on the wall. cAPs and wAPs are just ugly.
And sadly, MT has nothing to offer on the PoE switch side that comes near to ES and US (yet, I hope)...

Really?

But they are 2,4GHz only. Why not use the same external box, and stick a wAP AC (or its sucessor) inside?

I have not seen the new model for real, but the old cAP looks like a fire sensor. And sorry, but this one looks at first glance like an air intake of a ventilator with all those holes or dots. But anyway, much better than the old ones .


.

We use Ubiquiti AP with Mikrotik router because Ubiquiti does not have a decent router / firewall integrated with Wi-Fi AP in a single product.
Using the Ubiquiti firewall + AP is not competitively priced.

That's why you should use unifi when you are building home/office/hotel network: viewtopic.php?f=7&t=102908 If good wireless and its compatibility with all devices is not a priority for Mikrotik, then what is a priority and why we need products such as wAP or cAP?

Where does this stand now in 2019 after an entire 2018?

I've done both.

I've got many CAPsMAN installs with wAP AC's and cAP AC's in, without issue. I prefer Mikrotik for the flexibility and extra config options.

I had a big job that I had to put Ubiquiti in for the pure fact of short time frame and stock availability (this wasnt long after the factory fire) - I had a Mikrotik at the head-end still doing my hotspot, routing etc, but then ran a cloud key with about 30 Ubiquiti NanoHD's, didnt have any major problems. My gripe with Ubiqiuti is always the management, I dont like the cut down management portals, and I feel you can do less with them because of that. Given preference I will always tend to stick to all Mikrotik (except now i got 30-40 Ubiquitis to re-sell, as the initial job was only a temporary job)

Where does this stand now in 2019 after an entire 2018?

I think series of recently released products (and I'm not talking last few months but last few years) kinda draws line where is mtk focus. And "office" devices don't seem to be that. cAP and wAP seem to be only filler in offer for those who want to go full mikrotik. But general focus seems to be on backbone/core/edge networking - wired routers, switches, wireless point-to-point/CPE etc. Wireless in AP sense seems to be additional feature of mtk while UBNT is built around it from ground up. In their case wired networking is kind of "filler" because from what I've seen so far edge router series is meh and unifi SWG, especially XG model, as well as their absolutely unjustified shift into making... servers??? (with freaking Ubuntu, why am I not surprised lmao) and damn LED lighting (seriously? No, really...) kind of shows me that they have no fuckin idea what they're doing in this field. It looks like hardware targeted for people with Apple-like mind share who just bought APs and got astonished by how you can do everything with one button which costs only half of hidney so now they want to go full UBNT just because matte aluminum looks sooo 1337 in rack... Unfortunately YouTube shows that there's a lot of people like that (and yes they legitimately bring up argument that it looks good in rack) but on the other hand it may be just that this kind of people is more likely to upload videos on YT...

And if you think about it, it all kinda makes sense - people responsible for building backbone/core networking are usually quite competent, higher tier technicians unlike many people who are responsible for office wifi, who don't even like their job, just clock-in 8-16 (because it's usually considered waaaaaaaay less critical than core network thus employers skimp on salaries here. Of course I'm totally generalizing now and it's not always true, sometimes it's even that the same people responsible for both parts of infrastructure but I refer to one of common scenarios, don't take it as offence) and whatever people here are going to say: no, MikroTik is NOT user friendly, idiot proof device, and it's far cry from that comparing to UBNT with pretty, colorful buttons. MikroTik devices are entirely different from UBNT and ROS is entirely different OS, different philosophy, different priorities, different everything.

Sure UBNT die hards will say hurr durr SWG has CLI. Sure, Windows also has CLI. Do we really want to compare Linux terminal to Windows terminal? Because that's kind of similar ground to ROS CLI vs one in SWG. If you really want to use ROS up to its full potential you need to know how to script it. It already cuts off lots of networking technicians who are too lazy to learn that "because I'm not programmer". It's similar level of thinking to sys admins who don't want to learn shellscripts...

All in all it kind of boils down to argument similar to Linux vs Windows or vs OS X. It's argument about philosophy, hierarchy of priorities in life and attitude. Argument that has really few to do with actual technical aspects of those devices since it's comparing apples to oranges.

I always claimed that preferring ease of use over possibilities is a sign of weakness. Change my mind.

Indoor:

UniFi performs better than Capsman
UniFi Devices had better Range than WAPac and CAPac


Outdoor:

Mikrotik with ARM Based Chipset is completely unusable, unstable Connections, ping lags, disconnects, TDMA not working with ARM, no Spectral Scan

AirMax and other like Mimosa are miles away

With Spectral Scan you easily finds the best channel, or see the reason for bad speeds

TDMA working with good speeds an jitter free transfer for voip

Smal channel sizes like 10mhz are Optimum for Backup links for a 60Ghz Mikrotik link
They don’t care about the needs so we don’t buy it anymore for Wireless!

Indoor:

UniFi performs better than Capsman
UniFi Devices had better Range than WAPac and CAPac


Outdoor:

Mikrotik with ARM Based Chipset is completely unusable, unstable Connections, ping lags, disconnects, TDMA not working with ARM, no Spectral Scan

AirMax and other like Mimosa are miles away

With Spectral Scan you easily finds the best channel, or see the reason for bad speeds

TDMA working with good speeds an jitter free transfer for voip

Smal channel sizes like 10mhz are Optimum for Backup links for a 60Ghz Mikrotik link
They don’t care about the needs so we don’t buy it anymore for Wireless!

Don't get the angryness of these posts. Every vendor has limited resources and focus on stuff they see their market. Take the best of every vendor to build your network. I am not from US but I like their approach to praise the goodies and don't keep on blaming stuff that is not good.
MT has a lot of very good stuff to buy and a lot of stuff I don't buy. UBNT/Mimosa/Cambium has stuff I buy and stuff I don't buy.

Where does this stand now in 2019 after an entire 2018?

I feel like there's been no real progress since the original hAP AC release. I'm still using wAP AC units when I need a small cheap AP and don't care about latency, but for any big deployment I'm going with UBNT / Ruckus depending on budget.

Don't get the angryness of these posts. Every vendor has limited resources and focus on stuff they see their market. Take the best of every vendor to build your network. I am not from US but I like their approach to praise the goodies and don't keep on blaming stuff that is not good.
MT has a lot of very good stuff to buy and a lot of stuff I don't buy. UBNT/Mimosa/Cambium has stuff I buy and stuff I don't buy.

People are mad because they feel that something has been taken away from them. Older products didn't have issues with NV2 or TDMS etc., only new ARM devices are problematic from what I've seen so far. It's the same way I'm totally mad about RB4011 because it's hugely disappointing device in my opinion that is not actual RB2011/RB3011 successor since they took away a lot of features that were present in RB2011/RB3011. At this point fact that it has superior performance is less relevant than fact that it's missing features. Pretty much like ARM devices. Nobody cares that they perform better if they have inferior feature set.

People chooses UniFI/UBNT because

• Better looks
• Better range and performance
• Simplier and better looking WebGUI

And all above is true - but Mikrotik is far better at routing and more in depth functions.

what I've seen so far edge router series is meh

TBH, I absolutely love to work on the EdgeRouter's CLI, rarely look at the GUI at all. It's a Linux that you can script to your liking. It also has a feature (not out of the box but easily added) to forward broadcasts on certain ports into different subnets. This was a life saver for my home install. AFAIK this can not be added to ROS.

My 2 cents...

I choose Mikrotik:
- For routing when high performance-to-price ratio and advanced features are needed.
- When configuration by the user is not required.
- For the power of winbox.
- For Wireless when MkTik routers are also in use and vendor standardization is important.
- For the really good online documentation (which can always become better).
- Because it is a European company, which pioneered in and still leads the "SOHO router" market space while it has less access to funding than its US-based counterparts and for me, it is important to support them.

What I wish Mikrotik did differently:
- Put more care into their wireless products. After all, this is where they started from. Better noise filtering, TDMA performance, statistics, even tx-power settings, all cry for more attention.
- Invest in a unified and easy for the end user, management GUI on top of the great but nerd-oriented webfig/winbox.
- Extend their cloud with more features (like management).
- (Finally) introduce a new kernel into ROS.

<- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ->

I choose Ubiquiti:
- For a unifi(ed) management platform.
- When ease of configuration even by end-users is needed.
- For the beautiful, accessible (cloud-based) (but also browser-heavy) GUI with statistics.
- For much more complete wifi products and better wifi performance.
- For much better management of the PtP wireless and other devices (on the fly tx-power setting, dedicated radio for scanning and configuration, detailed statistics, UNMS).

What I wish Ubiquiti did differently:
- Put more features into the GUI, while keeping their intuitive approach.
- Have something as light and powerful as winbox is.
- Improve their PtP performance.
- Introduce 60GHz PtP and PtMP products.
- Write better documentation.

As a conclusion, I would say that ubnt is better in wireless performance and excels in ease of management/use, while MkTik excels in configurable options, features and routing performance.

My 2 cents...

I choose Mikrotik:
- For routing when high performance-to-price ratio and advanced features are needed.
- When configuration by the user is not required.
- For the power of winbox.
- For Wireless when MkTik routers are also in use and vendor standardization is important.
- For the really good online documentation (which can always become better).
- Because it is a European company, which pioneered in and still leads the "SOHO router" market space, it has less access to its US-based counterparts and for me, it is important to support them.

What I wish Mikrotik did differently:
- Put more care into their wireless products. After all, this is where they started from. Better noise filtering, TDMA performance, statistics, even tx-power settings, all cry for more attention.
- Invest in a unified and easy for the end user, management GUI on top of the great but nerd-oriented webfig/winbox.
- Extend their cloud with more features (like management).
- (Finally) introduce a new kernel into ROS.

<- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ->

I choose Ubiquiti:
- For a unifi(ed) management platform.
- When ease of configuration even by end-users is needed.
- For the beautiful, accessible (cloud-based) (but also browser-heavy) GUI with statistics.
- For much more complete wifi products and better wifi performance.
- For much better configuration of the PtP wireless devices (on the fly tx-power setting, dedicated radio for scanning and configuration, detailed statistics).

What I wish Ubiquiti did differently:
- Put more features into the GUI, while keeping their intuitive approach.
- Have something as light and powerful as winbox is.
- Improve their PtP performance.
- Introduce 60GHz PtP and PtMP products.
- Write better documantation.

Do all people asking for new kernel realize that it would mean dropping support for WHOLE current CCR series since Linux kernel officially dropped support for Tile-Gx CPUs architecture? While I'm not saying Tile-Gx is awesome it'd still mean dropping support for devices that are:
1) still being sold in fact
2) not even few years after EOS.
3) kinda flagship series of routers...

Regarding cloud management maybe it'd be usable in some less critical scenarios yet I personally don't know single client that would expose such critical infrastructure element as router to any form of cloud management. It's ridiculous. Especially in the age of questionable practices of institutions like MI6, NSA etc. You don't even need vulnerabilities. It's built-in backdoor on demand.

imho nerd oriented/powerful and simple/intuitive are exact opposite directions and you can't combine them. It'd require separate "user-friendly" UI option. I think QuickSet in webfig was supposed to be that but failed miserably.

I guess I agree with the rest of points.

Though I'd point out how UBNT fails to grasp concept of uplink interface in routers for router on the stick configs. That's something beyond me.

Though I'd point out how UBNT fails to grasp concept of uplink interface in routers for router on the stick configs. That's something beyond me.

I don't have any experience with Ubiquity, but I did use some other consumer-oriented devices. And I'd say that sticking to some (hard-coded?) concepts such as separate uplink interface (which kind of kills concept of router on a stick) is part of making those devices much more user-friendly than mikrotik.

And I don't agree that QuickSet failed miserably. Perhaps it'd need some more operation mode (such as switch+AP mode, but some modes are not supported by other vendors in the same price range, so why bother?) ... What might be proper way to go is perhaps detect that some setting on device was done outside QuickSet and as result to (permanently) disable QuickSet option for such device. My feeling (based on forum posts) is that quite some users changed a thing or two outside QuckSet, then returned to QuickSet and changed a thing or two there ... and this last step messed the whole config.

Though I'd point out how UBNT fails to grasp concept of uplink interface in routers for router on the stick configs. That's something beyond me.

I don't have any experience with Ubiquity, but I did use some other consumer-oriented devices. And I'd say that sticking to some (hard-coded?) concepts such as separate uplink interface (which kind of kills concept of router on a stick) is part of making those devices much more user-friendly than mikrotik.

And I don't agree that QuickSet failed miserably. Perhaps it'd need some more operation mode (such as switch+AP mode, but some modes are not supported by other vendors in the same price range, so why bother?) ... What might be proper way to go is perhaps detect that some setting on device was done outside QuickSet and as result to (permanently) disable QuickSet option for such device. My feeling (based on forum posts) is that quite some users changed a thing or two outside QuckSet, then returned to QuickSet and changed a thing or two there ... and this last step messed the whole config.

My problem with lack of uplink in devices is that I use for example two CRS317 switches attached to hypervisors and I need "somewhat 10G" router for inter-VLAN routing with statefull firewall between zones. Firewall is obviously crucial here since if I didn't need that additional security I could just stuff everything into single VLAN and call it a day. Stateless firewalls are pain in the ass to configure securely. There are jumbo frames everywhere so I don't necessarily need hardcore performance. With MikroTik my options start from CCR1009 for like 500$, go through CCR1016, CCR1036 and then there's somewhere CCR1072. With UBNT to get 10G port I have EdgeRouter™ Infinity for two kidneys. THE END. And I'll be using one or two 10G ports on it. 10/10.

Modes available in QuickSet depend on particular device. I use various MikroTik devices and they offer various predefined modes in QuickSet

Do all people asking for new kernel realize that it would mean dropping support for WHOLE current CCR series since Linux kernel officially dropped support for Tile-Gx CPUs architecture? While I'm not saying Tile-Gx is awesome it'd still mean dropping support for devices that are:
1) still being sold in fact
2) not even few years after EOS.
3) kinda flagship series of routers...

Good, TILE is a dead architecture anyway, RB4011 outperforms it on a core for core basis. Users with those devices can continue using old RouterOS version, it should not be a reason to hold development back on everything.

Mikrotik employees many times stated that they are not using built-in kernel module for TILE architecture. Instead, they are using their own module developed in cooperation with manufacturer of those CPU. Dropping TILE support from new kernel is not relevant to RouterOS.

Mikrotik employees many times stated that they are not using built-in kernel module for TILE architecture. Instead, they are using their own module developed in cooperation with manufacturer of those CPU. Dropping TILE support from new kernel is not relevant to RouterOS.

I was about to comment when I saw your post. Indeed they stated that, back when TILE support was dropped from the Linux kernel. But while this makes RouterOS more independent, it also means that they will have to re-develop the TILE support for a new kernel in-house, which is obviously resource-intensive and hard to debug, as the test base is far more limited.

On the other hand, I am in no way a kernel expert but I keep reading here that many of the current limitations/ shortcomings of ROS cannot be overcome (or are much difficult to overcome) with the current, aging kernel. This is why I have come to believe that the moment for a switch has come. But again, as said, I am no expert in this.

So it seems it's CCR versus ROS7 ... I wonder who's loosing?

If this is the case, then Mikrotik urgently needs to introduce new line of high-end routers which will replace current CCR roster, based on some modern platform (I wonder if ARM is that platform). I don't think Mikrotik can survive with top of the line routers new from factory running obsolete ROS version.

Only after that we'll see ROS7.

So it seems it's CCR versus ROS7 ... I wonder who's loosing?

If this is the case, then Mikrotik urgently needs to introduce new line of high-end routers which will replace current CCR roster, based on some modern platform (I wonder if ARM is that platform). I don't think Mikrotik can survive with top of the line routers new from factory running obsolete ROS version.

Only after that we'll see ROS7.

Considering that another Tile CCR has been released like few days ago I guess we're gonna wait xD. Yeah it probably will be ARM but I think MikroTik can't move to ARM on flagships yet considering how broken ARM version of ROS is as of now. RB4011 is basically bugged disaster.

So it seems it's CCR versus ROS7 ... I wonder who's loosing?

If this is the case, then Mikrotik urgently needs to introduce new line of high-end routers which will replace current CCR roster, based on some modern platform (I wonder if ARM is that platform). I don't think Mikrotik can survive with top of the line routers new from factory running obsolete ROS version.

Only after that we'll see ROS7.

Considering that another Tile CCR has been released like few days ago I guess we're gonna wait xD. Yeah it probably will be ARM but I think MikroTik can't move to ARM on flagships yet considering how broken ARM version of ROS is as of now. RB4011 is basically bugged disaster.

But all the complains about ARM are wireless related, right? A pure router (a true CCR) could do well, couldn't it? The 4011 has some problems with the FSP+ ports - but they are chipset related, not CPU related.

Or I am missing something?

But all the complains about ARM are wireless related, right? A pure router (a true CCR) could do well, couldn't it? The 4011 has some problems with the FSP+ ports - but they are chipset related, not CPU related.

Or I am missing something?

viewtopic.php?f=3&t=138613
Many minor issues like reboots, freezes, SFP issues plenty of butthurt. I guess ARM devices also don't support CPU clock adjustments. I also personally use hAP ac2 and I didn't manage to get USB modem to work. The same modem plugged to CCR1009 works flawlessly so I ended up connecting 3g backup line to core router instead of edge router which doesn't really make much sense tbh from infrastructure standpoint but it was the only option I had...

Basically ARM support is quite incomplete comparing to other platforms including Tile-Gx.

But all the complains about ARM are wireless related, right? A pure router (a true CCR) could do well, couldn't it? The 4011 has some problems with the FSP+ ports - but they are chipset related, not CPU related.

Or I am missing something?

viewtopic.php?f=3&t=138613
Many minor issues like reboots, freezes, SFP issues plenty of butthurt. I guess ARM devices also don't support CPU clock adjustments. I also personally use hAP ac2 and I didn't manage to get USB modem to work. The same modem plugged to CCR1009 works flawlessly so I ended up connecting 3g backup line to core router instead of edge router which doesn't really make much sense tbh from infrastructure standpoint but it was the only option I had...

Basically ARM support is quite incomplete comparing to other platforms including Tile-Gx.

Well, apart wifi (still a mess), almost all problems on this thread are related to SFP module on SFP+ cage needing auto negotiation disabled. I saw one (two?) people complaining about instability - but they gone away. Maybe a lemon? Many others said it was stable on production, so...

But all the complains about ARM are wireless related, right? A pure router (a true CCR) could do well, couldn't it? The 4011 has some problems with the FSP+ ports - but they are chipset related, not CPU related. Or I am missing something?

I'm loving the RB4011. I don't use the WiFi model.

I saw one (two?) people complaining about instability - but they gone away. Maybe a lemon?

I was complaining about stability of hAP ac² (also ARM based). The problem was freezing of all wired ports (I don't really use wireless on that unit so I don't know if wireless froze as well) and had to configure watchdog for it to automatically recover. I've since changed configuration (went from /interface ethernet switch based VLAN setup to /interface bridge based setup) and I'm running beta versions of ROS.

It runs stable now, not sure which change did the trick (possibly the move towards bridge vlan-filtering as switch chip has at least one known bug regarding VLANs). Which also hints that problems with ARM devices are actually problems with peripheral devices (switch chips, wireless adapters) rather than CPU platform itself.

Hence if MT introduced high-end router based on ARM platform and wouldn't save dimes on peripheral hardware, it might prove as a stable and potent device.

@mkx: Interesting! similar happened to me when I tried to limit bandwidth to one particular port via switch menu! Whole unit was disconnecting on regular basis.. I guess the switch in RBD52G is not that good after all

@mkx: Interesting! similar happened to me when I tried to limit bandwidth to one particular port via switch menu! Whole unit was disconnecting on regular basis.. I guess the switch in RBD52G is not that good after all

I'm using hardware switch a lot in hAP ac², even with 9k jumbo and looping traffic twice through that switch (from cpu port to 2x HA inline L2 IPS, back to hAP, to second router) and didn't experience instabilities. It sounds like coin toss. I'm very satisfied with hAP ac² switch because I use 9k jumbo in whole LAN so it was nice to be able to do all fragmentation on edge router.

That said I didn't limit bandwidth on switch.

When I first started to play with RBD52G, I used it as simple switch/AP without VLANs ... and it ran stable. So it seems that switch chip doesn't like any filtering (either bandwidth filtering or VLAN tag manipulation or ...).
As I reported in another thread a while ago, MT confitmed a switch chip bug regarding untagging PADO (or something) packets ... which affects running PPPoE on top of vlan interface over HW-offloaded access port. Not very common scenario, I know. The same bug doesn't happen if access port is on another device (I checked both RB951G and a D-Link managed switch), nor does it happen with bridge vlan-filtering. PPPoE client is running on RBD52G all the time.

@mkx: Interesting! similar happened to me when I tried to limit bandwidth to one particular port via switch menu! Whole unit was disconnecting on regular basis.. I guess the switch in RBD52G is not that good after all

I'm using hardware switch a lot in hAP ac², even with 9k jumbo and looping traffic twice through that switch (from cpu port to 2x HA inline L2 IPS, back to hAP, to second router) and didn't experience instabilities. It sounds like coin toss. I'm very satisfied with hAP ac² switch because I use 9k jumbo in whole LAN so it was nice to be able to do all fragmentation on edge router.

That said I didn't limit bandwidth on switch.

Well it didn't work for me. Cause seemed linked with wifi, after investigation with the support they asked me to return the unit. Had also some switch vlan configurations. I have no use for a unit that hang at least once a day. Maybe that was just bad luck but I am not willing to try out all the units of my supplier to tell...