Well, that would seem to be less than optimal from a marketing perspective.

 

“Custom chipset” or not, you have to suspect it follows that typical path of an 
analog front end SOC (A/D, D/A, PA, programmable filters, RF switch, etc.) and 
an FPGA/CPU chip.  Likely they went to one of the usual suspects like Analog 
Devices, Altera, etc. and had chips customized to their specs.  Or there are 
communications oriented fabless design houses that could do this.  My point 
being it is probably a programmable radio and could be backward compatible with 
airMax if they wanted.

 

I could be wrong, they could have looked at the programmable radio approach and 
decided it would be no different from the Canopy/450 approach and would not 
allow “disruptive pricing”.  And had a chipset made with dedicated hardware for 
the various OFDM functions like FFT, coding, clock recovery, etc.  Similar to 
how an 802.11 chip is designed, but without the advantage Mimosa had with 
Quantenna, leveraging the volume of mass market 802.11 products.  In that case, 
they might have thrown backward compatibility with their own products overboard 
to meet cost objectives.

 

Ubiquiti certainly has their challenges, they are fighting a war on two fronts 
– Mimosa and Cambium.

 

Another question for LTU is frequency bands.  We all understand that 802.11 
based radios will be limited to 2.4 and 5 GHz, unless you use frequency 
conversion, which never seems to work all that well.  But a custom approach 
like 450 or airFiber is not limited to WiFi bands.  What bands will Ubiquiti 
decide to go after, and how will they price it?  Look at AF3x, it demonstrates 
the flexibility of a custom platform, but it’s twice the price of AF5x.  It’s 
actually more expensive than Cambium’s 3 GHz PTP450.  So much for disruptive 
pricing.  But if they decide to cherry pick and only go after 5 GHz and maybe 
2.4 GHz, will you use another vendor for 900 MHz and 3 GHz in your network?  
Same question I raise regarding AF11x, or for that matter B11.  OK, that’s 
nice, I have an inexpensive 11 GHz backhaul product I can buy.  But I still 
need to buy my 6, 18 and 23 GHz backhauls from someone else like 
Ceragon/Cambium, SAIE, SAF, Trango, Dragonwave, Exalt, etc. that offers a full 
line.  I don’t like that, I’d rather pay a bit more for 11 GHz and be able to 
use the same vendor for all my licensed links.  But I also don’t like going to 
WalMart for milk, bread and diapers, Whole Foods for organic meat and 
vegetables, and a regular grocery store for everything else.  Other people will 
make the 3 trips.

 

 

From: Af [mailto:[email protected]] On Behalf Of Josh Reynolds
Sent: Thursday, December 8, 2016 9:33 AM
To: [email protected]
Subject: Re: [AFMUG] 450m , No News is Good News?

 

The AFLTU chipset is brand new with no legacy bits. Their 2nd real custom 
chipset made at UBNT. Most likely a forklift.

 

On Dec 8, 2016 9:24 AM, "Ken Hohhof" <[email protected] <mailto:[email protected]> 
> wrote:

AF software features will have to move lightyears ahead for LTU to compete with 
450/450i/450m.

AF is very barebones, kind of like a licensed backhaul, and can get away with 
it because it's just PTP.  Cambium has 15 years of adding software features and 
pulling all of them forward to the new platforms.  Sure, it's just a Small 
Matter Of Programming, but even now every time Cambium adds some features there 
are some bugs to work out, so trying to bring out a new PMP platform with all 
those features on day one would be a monumental undertaking.

Also, each PTP links is kind of a one-off, other than spares, training and 
network management systems.  You can put in an airFiber or Mimosa link, then 
decide to use something else for your next link, and that first link will 
continue working fine.  PMP is different (assuming proprietary systems), 
because once you deploy APs, all the CPEs have to match.  And with GPS sync, 
neighboring towers may have to use the same system.  It's a big decision.  Yes, 
the cost is also a big decision, but you don't want to deploy a bunch of PMP 
equipment and then end up hating it or finding the vendor didn't deliver on the 
"implemented in future firmware" promises.

Has anyone mentioned network transitions?  Elevate lets you transition an 
airMax network to ePMP.  Is it fair to assume LTU will do the same?  Or is it a 
forklift upgrade?


-----Original Message-----
From: Af [mailto:[email protected] <mailto:[email protected]> ] On Behalf 
Of Seth Mattinen
Sent: Thursday, December 8, 2016 8:47 AM
To: [email protected] <mailto:[email protected]> 
Subject: Re: [AFMUG] 450m , No News is Good News?

On 12/8/16 6:23 AM, Mike Hammett wrote:
> Don't get me wrong, I think LTU will be out next year and I do think
> it'll be a big hit, mitigating a lot of the issues we've seen in their
> AirMax platform. It has just seemed like it takes a lot to get any of
> the groups there to really do what the customers want.


Certainly AirMax seems to have some design limitations. Lowly ePMP can do sync 
on an Atheros-based platform while AirMax just can't seem to get there at the 
same level.

For me it's UBNT's ADD with abandoning things that makes me nervous.
Every vendor claims things that I'll believe it when it ships, but UBNT will 
release something, make some progress, then drop it because it's no longer new 
enough to raise stock prices or becomes unsexy long term maintenance. Like 
AirControl, then AirControl 2, then cloud-whatever, then AirCRM, now is it back 
to AirControl 2 again? I don't even know. I still have some M gear and ePMP 
Elevate is tempting just to get away from the AirControl dumpster fire.

~Seth



Reply via email to