Hi

I am just wondering. I have seen posts discussing RouterOS v7 since 2014 (atleast). It's 2018 now. That's 4 years of wait time for things like multi threaded BGP etc.

If it is because of financial constraints, I am sure a lot of Mikrotik users around the world would gladly contribute to the cause that we can accelerate the release of v7 by helping through crowd funding.

We could probably get a kickstarter or gofundme etc going if for a fact we knew that this is what's holding Mikrotik back.

Mikrotik isn't a startup...

Thanks for the concern. Maybe there is sarcasm in there, but no, MikroTik doesn't have money issues. MikroTik is a big company with a lot of history, RouterOS has been made for over 20 years and has become more and more complicated, so major versions are coming our slower than before.

I don't know which particular feature you don't have in v6, that you expect will exist in v7, but you have to be patient. Development takes time and resources. And when I say resources, I mean people-hours and brainpower. We are working on it, don't worry.

I don't know which particular feature you don't have in v6

OP clearly stated he is missing multi-threaded BGP. I guess it was promised?
If I should speak for myself, I would like to point out there are many LTE modems which are marked as "available in v7".

And when I say resources, I mean people-hours and brainpower

both people-hours and brainpower are equal to money. It is always possible to hire more people. It is always possible to hire higly skilled people. (and now I have absolutely no intention to suggest that your employees are not skilled. On the contrary. But there will always be some world-class genius who expects much higher salary) And it will always end up with money - how much can be spent on software development.

To sum up: I don't think anyone is worried that Mikrotik might have financial troubles. On the other hand, it would be naive to think that both HW and SW development wouldn't be faster if you somehow magically doubled your profit. Expenses are always limited by income, and every company has limited income. Thats simply how it works.

It might be easy to hire any kind of developer, but to find a person who can quickly adapt and start working on important v7 RouterOS features - not an easy task.

Anyway. Multi Threaded BGP doesn't exist. It will not be coming in v7 and is not implemented in any other brand routers. You can read other topics about it.
What we can promise, is simply a faster loading of the BGP routing table than in v6.

EDIT: here you go, about multithreaded BGP: viewtopic.php?f=3&t=129649&p=638204&hil ... re#p638204

I fully agree with your statement that it is not easy to quickly hire person who can jump a train. However, we both know that it would be easier with money, than without money

Anyway, this isn't really useful discussion. sorry for wasting your time. Certainly the idea of crowdfunding isn't really ideal. On the other hand - something like vote for "most wanted" functions would be cool. I always wonder how you guys decide which way to go, what things to focus on etc... hopefully its not like this

I always wonder how you guys decide which way to go, what things to focus on etc... hopefully its not like this

You owe me a new keyboard.
10 minutes laughing, and counting.

I don't know which particular feature you don't have in v6, that you expect will exist in v7

Can I start? I expect:

• spectral scan on AC hardware

• Tx power table on AC hardware

• better troughput on AC hardware

• new kernel and solving problems with ARM and NV2 protocol

• different channel widths on the AC hardware. For example, 10 and 30Mhz

• we have AC wave2 hardware, but we do not have software for it

and more...

I don't know which particular feature you don't have in v6, that you expect will exist in v7

Can I start? I expect:

• spectral scan on AC hardware

• Tx power table on AC hardware

• better troughput on AC hardware

• new kernel and solving problems with ARM and NV2 protocol

• different channel widths on the AC hardware. For example, 10 and 30Mhz

• we have AC wave2 hardware, but we do not have software for it

and more...

me too

Development takes time and resources. And when I say resources, I mean people-hours and brainpower. We are working on it, don't worry.

@normis: Personally I believe that most of the issues are due to the closed source and properitary nature of most of the components used within ROS. Instead of re-inventing the wheel with ROSv7, maybe it would be better to more-intensely rely on open source stuff such as OpenSSH, OpenVPN, FRRouting, etc. Also it might be beneficial to focus at one (or two) architectures like ARM64 and x86 in the future, instead of maintaining five architectures plus legacy hardware at once... As we can see, complexity hinders innovation

I think most uf users here are not missing features, but bugfixes.
Like IPv6 VRF (yes, that's a bug - a huge one), IPv6 recursive gateway lookup (I still don't understand why that should be kernel related - thats the job of a routing daemon, not of a kernel), faster BGP, ...

MikroTik still requires their job applicants to speak Latvian - maybe that's a problem too.

We need a update of routerOS kernel because we need a faster BGP routing.
Mikrotik should let a multithreading feature in the kernel.

Better Crowd Funding for cambium & fast port mikrotik hardware LHG/sxt mipsBE or any alternative software. MIkrotik wireless AC need separate from this company, they waste more money for kids kontroll & many other idiotic features Where spectral scan, NV2 for arm or another working TDMA protocol ?

Just for kicks, a paper about multithreaded BGP:

http://www.thinkmind.org/download.php?a ... 4_10_10029

Better Crowd Funding for cambium & fast port mikrotik hardware LHG/sxt mipsBE or any alternative software. MIkrotik wireless AC need separate from this company, they waste more money for kids kontroll & many other idiotic features Where spectral scan, NV2 for arm or another working TDMA protocol ?

+ 100

It might be easy to hire any kind of developer, but to find a person who can quickly adapt and start working on important v7 RouterOS features - not an easy task.

Anyway. Multi Threaded BGP doesn't exist. It will not be coming in v7 and is not implemented in any other brand routers. You can read other topics about it.
What we can promise, is simply a faster loading of the BGP routing table than in v6.

EDIT: here you go, about multithreaded BGP: viewtopic.php?f=3&t=129649&p=638204&hil ... re#p638204

normis, Can you please tell some news on progress of v7?
I mean there were years since start of it's development and we still wait blindly...
I'm not asking time when it will be done just rough % of it's completion(as developer you surely know at which point you are in your work ).

It might be easy to hire any kind of developer, but to find a person who can quickly adapt and start working on important v7 RouterOS features - not an easy task.

Anyway. Multi Threaded BGP doesn't exist. It will not be coming in v7 and is not implemented in any other brand routers. You can read other topics about it.
What we can promise, is simply a faster loading of the BGP routing table than in v6.

EDIT: here you go, about multithreaded BGP: viewtopic.php?f=3&t=129649&p=638204&hil ... re#p638204

normis, Can you please tell some news on progress of v7?
I mean there were years since start of it's development and we still wait blindly...
I'm not asking time when it will be done just rough % of it's completion(as developer you surely know at which point you are in your work ).

I can't speak for Normis but no doubt you will get the typical response. It's in process, it'll be done when it's done. What exactly is it you cannot live without under v6?

I mean there were years since start of it's development and we still wait blindly...

You say that, like there had been no new features in v6. In fact, most exciting stuff that was meant for v7, was already ported to v6.

We don't need exciting stuff, we need bugs fixed.
Like IPv6 recursive routing, usable BGP.

I just want fast BGP so I can load a couple of full tables and support more than 1 MPLS VRF without getting packet loss every time bgp updates.

Keep the control pane for BGP separate and prevent the forwarding pane from using the same cores

We don't need exciting stuff, we need bugs fixed.
Like IPv6 recursive routing, usable BGP.

- Spectral Scan
- new Wireless Chipset support
- Support for Wave 2 (Speeds near competitors)
- Support for 802.11ax
- Working TDMA

- Full IPv6 Support

I mean there were years since start of it's development and we still wait blindly...

You say that, like there had been no new features in v6. In fact, most exciting stuff that was meant for v7, was already ported to v6.

OpenVPN with UDP connections

Mikrotik might consider dividing their software into different releases like Cisco does with IOS, IOS-XR and the SMB stuff.
Requirements are simply too different. CCR as BGP edge router, an access point or a home router are simply not comparable.

Mikrotik might consider dividing their software into different releases like Cisco does with IOS, IOS-XR and the SMB stuff.

You mean, like RouterOS and SwOS?

Mikrotik might consider dividing their software into different releases like Cisco does with IOS, IOS-XR and the SMB stuff.

You mean, like RouterOS and SwOS?

Lol yeah...

lol, sure... because when I log into to IOS and i run some command.and then i go to another IOS and run exactly same command, the result is different because one IOS is on ASA (firewall) while second is on 887 (router).
gimme break please... whole idea of splitting resources and developing multiple systems is ridiculous.

I mean there were years since start of it's development and we still wait blindly...

You say that, like there had been no new features in v6. In fact, most exciting stuff that was meant for v7, was already ported to v6.

I appreciate all new features you added in v6 but still there are features i wait for ... that are planned for v7.
Like adding more queue types like codel and fq_codel (active queue managing).
Like more parallelization of load(firewall and queues) for CCR(36 cores).
Also implementing features into v6 does not show anything of progress in v7, right?

What do you mean with queue "parallelization"? Each parent queue already works on separate CPU core in v6.

What do you mean with queue "parallelization"? Each parent queue already works on separate CPU core in v6.

Really?Because when i asked support why cant my RB750Gr3 route more than 150mbit of traffic with queues and single core gets stuck at 100% while router all cores arent going even over 50%, i was told the opposite of what you say.

What do you mean with queue "parallelization"? Each parent queue already works on separate CPU core in v6.

I mean under big enough load on my 36-core CCR just several cores load up fully and others are almost free.
Maybe there is a way to parallelize load for firewall and queues(those are top consumption under load) so it will distribute much more evenly between cores to improve bandwidth?
Surely it will require some kernel-side changes(as both firewall and queues run in kernel space) so it can be for v7 with adding new queue types.

Really?Because when i asked support why cant my RB750Gr3 route more than 150mbit of traffic with queues and single core gets stuck at 100% while router all cores arent going even over 50%, i was told the opposite of what you say.

What are your queues? Telepathists are on another forum.

What are your queues? Telepathists are on another forum.

good point

New Intel LAN card driver (for example i210) should be available on v7? (because kernel was updated)

Or v6.x will updated Intel LAN driver?

IPv6 routing mark (policy route) was promise on v7 since few years ago...