Community discussions

MikroTik App
 
syadnom
Forum Veteran
Forum Veteran
Topic Author
Posts: 794
Joined: Thu Jan 27, 2011 7:29 am

60Ghz future, can we get some leaked info?

Mon Dec 05, 2022 8:51 pm

So I like the new cube pro models, they perform nice and I use them in some niche setups. That said... they live in niche setups. They are missing some features or there are some product holes. I'll try to conherently describe groups of features I'd like to see.


AP/AP mesh links. It takes quite a few Cube SA to fill out a pop, and then need another 1+ for backhauls. A mesh link option would really help. It's difficult to put a monstrocity on someone's roof. 4-6 cube-sa are pretty clean looking but add in another layer of links for backhauls and it gets ugly fast... also, the effort in revisiting homes to put additional links in is a pain.
so the question is, do you have some mesh capabilities in the software pipeline?

Wider APs. 90s for example. This is a two fold issue, one is that I need a lot of Cube SA to build a pop, but I also need a lot of ethernet ports to power them. A 90 degree Cube SA would actually pair decently with a powerbox pro. There is no 6 port powerbox pro, we shoot up to netpowers with are physically large for this task. Yes, the narrower beam gets more range, but I'm not always in need of that. I'm deploying cnwave successfully right now and it's slightly less range than my math says Cube SA 90's would be. It's just a far simpler and easier to setup arrangement AND with slightly less gain there's less interference in a meshy type setup. Then there's the 5Ghz backup, again, many 5Ghz APs isn't good, can't re-use bands so it's a LOT of 20Mhz channels to chew up for backup.

Terragraph. ie, mesh connections, same channel on each AP, beamforming. This might be a big ask right now but it makes a lot of things simpler. I don't need the routing engine here, just the radio links brought in as individual interfaces. the routing would also be nice, but there are other solutions to that white there really aren't for a mesh build. ie IPv6+OSPF works really well for auto-forming a mesh on discreet 60Ghz interfaces and since the likelyhood of being at top modulation is good, shortest path is best 99% of the time. static interface costs can be added as needed.

Kind of a slightly off topic but relevant item, EoIP and/or GRE hardware accell on IPv6. This is likely a prerequisite for terragraph anyway (likely GRE tunnels to core) and for a mesh setup it works a lot better to tunnel customers from their CPE port to a bridge on the core. There are a number of ways to do this, MPLS+VPLS for example, or even vxlan, but I don't really want a big bridged network. IPv6+OSPF+EoIP to the core is a very nice setup for this that is within existing wISP knowledge to deploy. vxlan over ipv6 also would be nice. I don't think the CPUs on existing CPE are enough to handle fast connections in software right now.


So, mikrotik guys, are my thought process in line at all with how you are moving forward with this product and if so, can you leak a little information for us?

Who is online

Users browsing this forum: No registered users and 14 guests