Reply
Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync

8.1 betas on a certain forum here are for Gen2 hardware, but 8.1 in general is for Fixed Frame with GPS timing discipline

Emerging Member
Posts: 99
Registered: ‎07-30-2015
Kudos: 37
Solutions: 2

Re: Rocket 5AC PRISM GPS sync

I don't understand all this hardware... Certain other vendors have basically 3 CPE radios, a small Nanostation-like CPE, a 25 dbi Dish, and a Connectorized radio.... And 2 versions of the AP - one like a Prism, the other like an AC LIte.  Other vendors even have less radio options than that.

 

If it were me, I'd have 4 dishes: 300, 400, 500, 620.  Inside each of those dishes, you could put a AC Feed, an M2 Feed, M5 Feed or a Connectorized feed.  All dishes and feeds are stocked and sold separately.  Need to upgrade / downgrade?  Swap feeds.  Need a bigger dish?  Swap dishes.  

 

Bonus points:  AC Prism Feed and Airfiber 2X/3X/5X feeds.

Extra points: Feed can be inserted standard or slant 45.

 

I get it- everyone likes to fill cargo containers and warehouses.  But we need less products and more technology - AC Prism when it first was announced was like a dream come true - dynamic filtering, GPS sync support, but we're already going to Gen 2 AC and no GPS sync is here yet.

 

Yet vendor C has managed to take all of our 5 year old radios and put GPS sync into it... How bow dah?

 

 

 

 

Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync

The EPMP method for Fixed Frame on XM, XW is basically a refined version of the old XM airSync. The methods are very similar. Ubiquiti didn't develop airSync as far as Cambium did their own. Ubiquiti has continued to pay the price in WISP mind-share ever since. (As it turns out, Ubiquiti M gear outsells EPMP worldwide by a huge amount, so maybe they didn't care so much initially... But, the writing is on the wall!!)

 

This internal pressure pays out for you. The 5AC method for Fixed Frame is faster, can transport more data in less time, can handle errors more quickly, it is more sophisticated than EPMP in several ways...and due to this additional horsepower required, it may or may not work well on XM. Once XC and WA are finished (soon), then we'll get XW fixed frame, and hopefully after that, XM. 

 

The amount of time that this fixed frame implementation is taking is not huge, and the stability of early implementations is high. Ubiquiti is really doing a nice job here and I think you will appreciate it. They are testing for even the smallest abberations very early, and rapidly iterating towards GPS-disciplined Fixed Frame Nirvana. 

 

Cash me ousside!! How bout dah?

Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync


directconnect wrote:

If it were me, I'd have 4 dishes: 300, 400, 500, 620.  Inside each of those dishes, you could put a AC Feed, an M2 Feed, M5 Feed or a Connectorized feed.  All dishes and feeds are stocked and sold separately.  Need to upgrade / downgrade?  Swap feeds.  Need a bigger dish?  Swap dishes.  


I like your style!! 

Emerging Member
Posts: 141
Registered: ‎03-05-2013
Kudos: 4

Re: Rocket 5AC PRISM GPS sync

I think is that they plan with the IsoStation 5AC
Emerging Member
Posts: 74
Registered: ‎01-12-2009
Kudos: 9

Re: Rocket 5AC PRISM GPS sync

I don't think ePMP's sync is really very much like airSync at all... I'm not sure there's really anything else like airSync (for good reason... ). ePMP's behaves much more like a traditional Canopy-style fixed frame system than the crazieness that is airsync.

 

How ePMP sync compares to the AC fixed frame sync, I have no idea though.

Highlighted
Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync

EPMP's sync technique is similar to airsync, but refined with bug fixes and improvements past where airsync got...

Regular Member
Posts: 400
Registered: ‎10-11-2013
Kudos: 111
Solutions: 2

Re: Rocket 5AC PRISM GPS sync


MHoward wrote:

I don't think ePMP's sync is really very much like airSync at all... I'm not sure there's really anything else like airSync (for good reason... ). ePMP's behaves much more like a traditional Canopy-style fixed frame system than the crazieness that is airsync.

 

How ePMP sync compares to the AC fixed frame sync, I have no idea though.


TDD synced system should behave quite similar. A timeframe for downstream, one for upstream and gaps between. AP tells cpes via a mask when they are allowed to send.

Difference is the quality, modulations and may be some protocol details.

As 5AC has higher modulations and more cpu power than epmp it should work better if SW works. 

Emerging Member
Posts: 74
Registered: ‎01-12-2009
Kudos: 9

Re: Rocket 5AC PRISM GPS sync

ePMP uses fixed frames... if I remember right, airsync was supposed to be a dynamic system... as I recall, it also required communication between the synced APs - ePMP has none of that, as long as the settings match, it works. I don't really see the similarities. Maybe there are similarities in the underlying workings, but on the surface, they seem to be very, very different. Not to mention the ~50% throughput penalty for using airsync... (ePMP definitely doesn't have that issue). 

SuperUser
Posts: 9,764
Registered: ‎12-08-2008
Kudos: 6852
Solutions: 422
Contributions: 1

Re: Rocket 5AC PRISM GPS sync

[ Edited ]

But fixed-frame systems do have a performance penalty.   You can't have it both ways.   Witness the AirFiber duty cycle setting - with fixed frames you can't get any smaller frames than the fixed size.   You can vary the frame size (within reason) but it's really hard to do that dynamically without a lot of communication overhead between the units.    And as to intercommunication between the APs, in the AirSync system that was used to communicate which timings they were all using, not to do any of the timing/framing itself.   With AF that isn't even necessary, since the frame boundaries are all on pre-determined time ticks controlled by the GPS regulated clocks.

 

All I can say is that I do expect AC GPS Sync will work and work quite well once it's released.

Jim

" How can anyone trust Scientists? If new evidence comes along, they change their minds! " Politician's joke (sort of...)

"Humans are allergic to change..They love to say, ‘We’ve always done it this way.’ I try to fight that. "Admiral Grace Hopper, USN, Computer Scientist
Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync


MHoward wrote:

ePMP uses fixed frames... if I remember right, airsync was supposed to be a dynamic system... as I recall, it also required communication between the synced APs - ePMP has none of that, as long as the settings match, it works. I don't really see the similarities. Maybe there are similarities in the underlying workings, but on the surface, they seem to be very, very different. Not to mention the ~50% throughput penalty for using airsync... (ePMP definitely doesn't have that issue). 


AirSync is a fixed frame system with adaptive, automatic upstream/downstream percentage adjustment. It's still fixed in the sense that the behaviour across all APs is deterministic (even though they can all change downlink percentage at the same time). The way that airSync achieves this determinstic behaviour is quite similar to how EPMP does Fixed Frame. And they're both far behind 5AC Fixed Frame in several ways... I think you guys will be quite pleasantly surprised at how the 5AC fixed frame performs at 6X mode, compared to EPMP at 6X... 

New Member
Posts: 29
Registered: ‎10-11-2007
Kudos: 14
Solutions: 1

Re: Rocket 5AC PRISM GPS sync


MHoward wrote:

ePMP uses fixed frames... if I remember right, airsync was supposed to be a dynamic system... as I recall, it also required communication between the synced APs - ePMP has none of that, as long as the settings match, it works. I don't really see the similarities. Maybe there are similarities in the underlying workings, but on the surface, they seem to be very, very different. Not to mention the ~50% throughput penalty for using airsync... (ePMP definitely doesn't have that issue). 


ePMP can do fixed (75/25, 50/50, or 30/70) or flexible frames, along with 2.5ms or 5ms frame lengths.

 

In regards to airSync, I'm not sure what twinkletoes is talking about... maybe you're thinking of the airSelect feature on older firmware revisions?

 

"airSelect is a technology that avoids interference and increases throughput. It dynamically changes the wireless channel by periodically hopping to the least used channel in the Frequency List (user-defined) within a designated time interval (user-defined in milliseconds). airSelect tracks interference levels on each channel used, hopping more frequently to those with the least amount of interference."

Established Member
Posts: 2,251
Registered: ‎07-03-2008
Kudos: 476
Solutions: 32

Re: Rocket 5AC PRISM GPS sync

airSync was the original UBNT attempt at GPS-synchronized fixed frame operation using commodity WiFi silicon.  It almost worked.

Established Member
Posts: 2,251
Registered: ‎07-03-2008
Kudos: 476
Solutions: 32

Re: Rocket 5AC PRISM GPS sync


eejimm wrote:

But fixed-frame systems do have a performance penalty.   You can't have it both ways.


True!

 

Synchronous operation brings predictability -- at the cost of fairly low channel utilization.  Best-effort packet-based systems maximize channel fill -- but as the channel approaches fullness (or if even one bad actor shows up)  weird stuff starts to happen.

Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync


webformix wrote:

MHoward wrote:

ePMP uses fixed frames... if I remember right, airsync was supposed to be a dynamic system... as I recall, it also required communication between the synced APs - ePMP has none of that, as long as the settings match, it works. I don't really see the similarities. Maybe there are similarities in the underlying workings, but on the surface, they seem to be very, very different. Not to mention the ~50% throughput penalty for using airsync... (ePMP definitely doesn't have that issue). 


ePMP can do fixed (75/25, 50/50, or 30/70) or flexible frames, along with 2.5ms or 5ms frame lengths.

 

In regards to airSync, I'm not sure what twinkletoes is talking about... maybe you're thinking of the airSelect feature on older firmware revisions?


Uhh, well, I'm talking about airSync. It's Fixed Frame for M series. The development on it was stopped too early. It is implemented largely using the same mechanisms as EPMP fixed frame. Its use of the Atheros wifi chip is very similar. This knowledge is derived from actual analysis of both systems. The exception is that Cambium perfected it (as much as possible, it's not a magic bullet), where Ubiquiti quit early. This really was a problem for Ubiquiti, who lost some mindshare in the US market to Cambium. They know this and they aren't interested in repeating it.

 

Ubiquiti's Fixed Frame for AC series does a number of tricks to work smarter than EPMP. The scheduler is very sophisticated. At 6X it already performs much better than EPMP. As it is back ported to XW and then XM, I think EPMP Elevate will be less attractive at $30 a piece. Maybe Ubiquiti will release some free firmware to Elevate EPMP users to 5AC!

 

Ubiquiti will win back your business. I hear you need the cost savings these days, anyways Man Happy

New Member
Posts: 29
Registered: ‎10-11-2007
Kudos: 14
Solutions: 1

Re: Rocket 5AC PRISM GPS sync

I should have been more verbose in my post... I totally forgot about airSync... mostly because I never used it, and from what I've read in various forums, the feedback from users has been... uhhh not good. Secondy, the rest of that post was directed towards MHoward in regards to airSelect.... I think he was getting airSync and airSelect confused.

 

I'm a big fan of using the right tool for the job.... sometimes that's a Cambium tool, sometimes that's a Ubiquiti tool. I have a new network to roll into my own, and I thought maybe ePMP Elevate would be a good tool for this, but after much consideration, I can't use that tool. Problem is two fold... airgrids and compliance test mode (I'm surprised at how many WISP's have painted themselves into a corner with compliance test mode)... I'm going to leave it at that, but basically can't use ePMP elevate and UBNT AC mixed-mode+ is looking better and better, especially now that it's out of beta. So it's a new network (to me) and a new area, and a new day... so why not try something new... AirmaxAC Prism's + Horns and mixed mode... let's see what we can do here!

 

ps... I know this seems like rambling off-topic stuff to everyone, but twinkletoes and I know each IRL and have a storied past.

Established Member
Posts: 1,557
Registered: ‎05-20-2008
Kudos: 374
Solutions: 5

Re: Rocket 5AC PRISM GPS sync

We are comepting here with a Mikrotik WISP (around 15000 users) and a Cambium WISP with 10000 clients.
We are the only WISP left using UBNT still.
Mikrotik WISP is getting ready for Sync with Elevate when it is ready.
Cambium WISP already converted the leftover UBNT equipment and everything to Elavate + enjoying the fixed frame.
Only we are left in the dark.
All I hear is XM is so weak, it cant have fixed frame !
"XW will be attemted."
Really tired of hearing excuses while somebody else is already working fixed mode with XM and XW .
Maybe it is time for UBNT to license the Cambium s code and be done with it and move on...
Regular Member
Posts: 400
Registered: ‎10-11-2013
Kudos: 111
Solutions: 2

Re: Rocket 5AC PRISM GPS sync


doush wrote:
We are comepting here with a Mikrotik WISP (around 15000 users) and a Cambium WISP with 10000 clients.
We are the only WISP left using UBNT still.
Mikrotik WISP is getting ready for Sync with Elevate when it is ready.
Cambium WISP already converted the leftover UBNT equipment and everything to Elavate + enjoying the fixed frame.
Only we are left in the dark.
All I hear is XM is so weak, it cant have fixed frame !
"XW will be attemted."
Really tired of hearing excuses while somebody else is already working fixed mode with XM and XW .
Maybe it is time for UBNT to license the Cambium s code and be done with it and move on...


doush wrote:
We are comepting here with a Mikrotik WISP (around 15000 users) and a Cambium WISP with 10000 clients.
We are the only WISP left using UBNT still.
Mikrotik WISP is getting ready for Sync with Elevate when it is ready.
Cambium WISP already converted the leftover UBNT equipment and everything to Elavate + enjoying the fixed frame.
Only we are left in the dark.
All I hear is XM is so weak, it cant have fixed frame !
"XW will be attemted."
Really tired of hearing excuses while somebody else is already working fixed mode with XM and XW .
Maybe it is time for UBNT to license the Cambium s code and be done with it and move on...

We have a lot of MT gear and suffer from wireless part which is not updated for years now. Nv2 gives bad tcp speedtest results. We are bored with customer calling as their speedtest does not match his package.

We migrate to ubnt ac. Even without sync we see very good results with rocket ac ptmp.

We tested epmp 1000 and had problems with interference and had a lot firmware problems. 

So we are happy with ubnt ac now and will see what gps sync brings.

For sure gps sync will only improve installations where spectrum is a problem. With enough free spectrum the fixed up/down ratio will reduce performance as it does with other gps synced systems.

 

Epmp is still old 11n and so is Elevate. 

Established Member
Posts: 906
Registered: ‎04-30-2009
Kudos: 151
Solutions: 4

Re: Rocket 5AC PRISM GPS sync

EPMP sync is nothing like AirSync, I don't see how you can even type that with a straight face.  Airsync never came close to working and ubiquiti sold a bunch of hardware continually telling their loyal customers SOON!

 EPMP works!

TwinkleToes, when did you become such a shill?

 

The problem with ubiquiti is honesty.  With epmp we know what we are getting and everything works as advertised.  If you ever have a problem it gets addressed and they are very helpful( no finger pointing)

 

With ubiquiti they are always talking about how great it is or how great it is going to be and if you ever have a problem then you must be doing something wrong.   I use alot of ubiquiti and I learned long ago that it's very difficult to get true support with ubiquiti issues( and there has been alot of them through the years).  

Their equipment is inexpensive and it looks nice but don't ever assume it is going to work as advertised.

 

To sit here and listen to a comparison between ubiquitis unreleased alpha product and cambiums 3 year old stable working product is almost comical.  

How bout when ubiquiti has released a stable working gps AP then you can 'cash me outside'.  See you in a couple of years.

Veteran Member
Posts: 4,813
Registered: ‎08-26-2009
Kudos: 1358
Solutions: 39

Re: Rocket 5AC PRISM GPS sync

Dude, i'm not saying that airSync ever worked as well as EPMP FF. I'm saying that it is the technical equivalent, down to how the chip is used to create the effect of fixed frames. I can tell you that Ubiquiti did in fact quit TOO EARLY and they regret this decision now. It could have worked at a level similar to EPMP if this happened some other way...

 

Two years? It's already up and running today. And it can do things that are not possible on XM XW EPMP platforms, today. Nobody ever argues that airsync sucks. It sucks. Nobody ever argues that EPMP FF upload or latency sucks. It still sucks. It will always suck. AC FF does not suck in any direction. Once the few known issues are worked out, you'll see it in the BETA PIPELINE and then you can make the choice for yourself.

 

Anyways, shill? A shill is a friend of a fraudster who pretends to be an enthusiastic, legitimate customer. To be fair to me, i'm actually just an enthusiastic, legitimate customer of Ubiquiti. There's no fraud here. To top of off, I'm not lying. You can trawl through my posting history here to see how much of a "shill" I was regarding airSync.

 

If I was a shill, I couldn't post this example of fixed frame in production. Here it is, in production, on a 20mhz link, with ARQ turned completely OFF. No packet loss even, that's cool... (SHILL!!! FRAUD!!!)

 

PING 172.20.56.5 (172.20.56.5): 56 data bytes
64 bytes from 172.20.56.5: seq=0 ttl=64 time=10.858 ms
64 bytes from 172.20.56.5: seq=1 ttl=64 time=10.776 ms
64 bytes from 172.20.56.5: seq=2 ttl=64 time=11.200 ms
64 bytes from 172.20.56.5: seq=3 ttl=64 time=14.463 ms
64 bytes from 172.20.56.5: seq=4 ttl=64 time=10.494 ms
64 bytes from 172.20.56.5: seq=5 ttl=64 time=14.236 ms
64 bytes from 172.20.56.5: seq=6 ttl=64 time=14.264 ms
64 bytes from 172.20.56.5: seq=7 ttl=64 time=10.838 ms
64 bytes from 172.20.56.5: seq=8 ttl=64 time=13.911 ms
64 bytes from 172.20.56.5: seq=9 ttl=64 time=9.848 ms
64 bytes from 172.20.56.5: seq=10 ttl=64 time=9.850 ms
64 bytes from 172.20.56.5: seq=11 ttl=64 time=9.607 ms
64 bytes from 172.20.56.5: seq=12 ttl=64 time=9.345 ms
64 bytes from 172.20.56.5: seq=13 ttl=64 time=13.318 ms
64 bytes from 172.20.56.5: seq=14 ttl=64 time=9.158 ms
64 bytes from 172.20.56.5: seq=15 ttl=64 time=9.293 ms
64 bytes from 172.20.56.5: seq=16 ttl=64 time=8.883 ms
64 bytes from 172.20.56.5: seq=17 ttl=64 time=8.850 ms
64 bytes from 172.20.56.5: seq=18 ttl=64 time=8.688 ms
64 bytes from 172.20.56.5: seq=19 ttl=64 time=8.785 ms
64 bytes from 172.20.56.5: seq=20 ttl=64 time=8.439 ms
64 bytes from 172.20.56.5: seq=21 ttl=64 time=8.342 ms
64 bytes from 172.20.56.5: seq=22 ttl=64 time=8.169 ms
64 bytes from 172.20.56.5: seq=23 ttl=64 time=11.807 ms
64 bytes from 172.20.56.5: seq=24 ttl=64 time=11.560 ms
64 bytes from 172.20.56.5: seq=25 ttl=64 time=11.384 ms
64 bytes from 172.20.56.5: seq=26 ttl=64 time=11.464 ms
64 bytes from 172.20.56.5: seq=27 ttl=64 time=11.163 ms
^C
--- 172.20.56.5 ping statistics ---
28 packets transmitted, 28 packets received, 0% packet loss
round-trip min/avg/max = 8.169/10.678/14.463 ms
Reply