I would appreciate some experienced feedback on how to deploy a wireless backbone to bridge a few wired networks in a remote location where it is not possible to pull cable.
As seen in the attached image, the idea is to set up a link from the uplink location to a tower close to the remote locations, and from this tower set up separate links to each area. We assume only PtP links (no PtMP).

All wireless links would be around 10km, and we need up to 100Mbps between each location and the uplink. There will be IPTV multicast originating from the uplink, as well as some VoIP in the network.

Is this a normal way to solve this situation? Would DynaDish be a good choice for the links?

Thanks!

I'm using DynaDish5 for MANY tower backup links, they work great, and at LONG distances! I have some that are 20-30 miles.

BEWARE, they are VERY sensitive to noise, and will have low rates if you don't separate them physically and frequency. Today, I had a link that was slowing it down, that was almost 4 feet below in vertical separation and still wasn't enough.

I think because of the high-output TX power these have, they throw out a lot of noise to co-located radios, I wish Mikrotik would've put a ISO shield band around the side of the dish to help deflect other 5Ghz noise.

Also I've found that using Nstreme on AC will provide the most throughput. Witnessed up to 80-90% drop in throughput capacity on NV2 protocol, its just HORRIBLE with AC!!

I think because of the high-output TX power these have, they throw out a lot of noise to co-located radios, I wish Mikrotik would've put a ISO shield band around the side of the dish to help deflect other 5Ghz noise.

Agree. I found it helps to reduce the TxPower on adjacent radios, if their links will allow. Tweaking TxPower on current generation MT gear is something of a mystery shrouded in an enigma unfortunately...

Thanks for the warning.
As we may need 4-5 dishes in a single co-located tower, the interference may become a problem then.
I wonder if it is worth the extra cost to use Netmetal5+mANT30+Sleeve30 instead of DynaDish in those cases..?

Do you also have experience like our case with two wireless links in series between the networks, for both normal Internet traffic and multicast on the same links? I suppose it would work in theory, but it would be nice to hear from someone who already has a similar config in production..

Using the mANT30 will definitely bring several benefits:

- allow you to lower tx power thanks to the higher gain, thus reducing co-location interference
- beam width: look at the beam patterns of each at routerboard.com, mANT30 is a very directive antena, with much, much less "throwback", and narrower beamwidth, so in addition of irradiating less, it will also pick less interference.

An advice: buy the -PA (precision adjustment) version. Its a very high quality and directive antenna (2,5degrees beamwidth), installers used to mount just panel antennas usually struggle to properly align them.

Multicast requires additional tunneling/engineering to carry the IPTV traffic, so that the final device transmitting over the air to end user/location doesn't actually use multicast, as the radio will switch to the lowest MCS rate when TX'ing multicast packets.