x
Ultra-Broadband

Moto Finishes Two XG PON Trials

Motorola Inc. (NYSE: MOT) is quite proud to say it's been involved in two trials of XG PON, but that doesn't mean the 10-Gbit/s technology is going to be used any time soon.

In June, the company announced it had done an XG PON trial with Verizon Communications Inc. (NYSE: VZ), which had also tested the technology on Huawei Technologies Co. Ltd. gear. (See Verizon Tests XG-PON With Moto Kit and Verizon: Huawei No Shoo-In for XG PON .)

And Motorola has completed another XG PON trial with "a major Chinese operator," says Floyd Wagoner, Moto's director of product marketing.

That trial, in particular, is a sign that Motorola can give Huawei some stiff competition in PON. That's true even at the GPON level, Wagoner says. "We are moving forward in the Chinese market -- Huawei's market -- with price points that are better than theirs on ONTs in some cases."

Neither carrier is all that close to deploying XG PON, though. The China trial, for example, was a test of whether XG PON and GPON can share the same fiber, but the carrier has no immediate plan to actually use XG PON, Wagoner says.

The standard for XG PON is due to be ratified later this year, and most vendors seem to expect deployments won't start for three or four more years.

The first, obvious reason is the cost. Prototype, pre-standard XG PON gear costs 20 times what GPON would, Wagoner says. Even in three or four years, after the standard has been in place and equipment costs have dropped, XG PON could still be four times the cost of GPON, he says.

But there's also no urgent demand for XG PON. Verizon and other carriers think that regular GPON can handle the expected increases in bandwidth demand, Wagoner says.

XG PON is more about reaching larger numbers of subscribers -- feeding apartment complexes (especially in crowded Asian cities) and business parks. "It's not necessarily about getting more bandwidth to the average subscriber," he says. "When that changes, then you'll see the shift to the next tier."

Along those lines, carriers are interested in increasing PON's split -- the number of subscribers who share a central-office port -- to a 1:64 or even 1:128 ratio from today's 1:32, he says. And that brings up an interesting possibility. If a particular area has lots of 1:64 PON deployed and needs more speed, the carrier will probably ratchet the split back to 1:32 rather than move to XG PON, Wagoner says.

This partially explains why Motorola hasn't latched onto WDM-PON, which gives each subscriber a dedicated wavelength. "We don't see the market going there at this point," Wagoner says.

— Craig Matsumoto, West Coast Editor, Light Reading

Pete Baldwin 12/5/2012 | 4:26:10 PM
re: Moto Finishes Two XG PON Trials

The true appeal of 10G PON has been to increase the number of users, not the speed per user.  That's something I've been told for a while.  Even so, I'm surprised that the urgency for faster PON hasn't really kicked in yet. I guess the big rampup just isn't that close -- or they aren't feeling the pain of it yet.

ieeefellow 12/5/2012 | 4:26:02 PM
re: Moto Finishes Two XG PON Trials

Re: The first, obvious reason is the cost. Prototype, pre-standard XG PON gear costs 20 times what GPON would,


 


what's the current price ratio for 1G PON as compared to 100M PON?


Pete Baldwin 12/5/2012 | 4:25:55 PM
re: Moto Finishes Two XG PON Trials

I don't have an immediate answer, but it's got to be more like 4:1 by now. Anybody know for sure?

Pete Baldwin 12/5/2012 | 4:25:53 PM
re: Moto Finishes Two XG PON Trials

I'd agree with your point. I don't sense a lot of hunger and desperation around XG-PON, and the development that's required sounds expensive.


We'll continue to keep track of its progress, though.  It's always interesting to keep tabs on the next generation of technology.

paolo.franzoi 12/5/2012 | 4:25:53 PM
re: Moto Finishes Two XG PON Trials

 


Duh,


One thing you left out, at least in the model of deployment that FiOS uses is that much of the cost has almost nothing to do with the ONT electronics and optics.  The power and mechanical schemes that they have specified are not cheap on its own.  Gel filled connectors cost what they cost and they don't really have any downward pressure on them


Where I am heading with that is that once the cost of the manufacturing associated with GPON optics was under control, the ONT cost for a BPON ONT was about the same as that for a GPON ONT.


I also would think a 10G down/2.5G up would be a reasonable compromise.  The thing I have yet to figure out is if they are really willing to pay for the aggregation layer routing and switching upgrade.


seven


 

Duh! 12/5/2012 | 4:25:53 PM
re: Moto Finishes Two XG PON Trials

1) On pricing in our industry, those who talk do not know, and those who know will not talk.  


2) Any relation between cost and price is pure coincidence.


3) Do not attempt to extrapolate incremental cost for X-GPON over GPON from incremental cost from GPON over BPON.  The engineering and manufacturing challenges are different:


BPON to GPON transition required volume production of low cost 2.5 Gb/s capable APDs and 1.2 Gb/s DFB lasers.  This was more of a volume manufacturing challenge than anything else.    OLT burst receivers were a  technical challenge, but more to the point it was difficult to justify developing these low volume devices.  It's amazing what the promise of volume, plus a lot of arm twisting, can accomplish.


GPON to X-GPON is going to require volume production of low cost 10G APDs.  Apparently, yields of these devices are low, due to a very significant and intractible process challenge. The option does exist in G.987.2 to rebalance the power budget to use high launch power, PIN detectors and FEC instead of APDs.   The new band plan requires more complex filters in both the ONTs and the CO.  On the OLT side,  volume production of 1577 nm EML (or CML) transmitters will be needed;  I understand this to be a modest challenge in fabrication.  Also, existing GPON OLT burst receivers tend to be marginal at 2.5Gb/s and will need to be respun.  And, of course, higher speed FPGAs (or ASICs) will be needed on both sides.


But, more to the point, unless and until a big operator is prepared to put an XG-PON  RFP with big volumes on the table, there will be no justification at either the component, module or system level for the incremental development and NPI.

Duh! 12/5/2012 | 4:25:50 PM
re: Moto Finishes Two XG PON Trials

Seven,


I was more focused on differences than equivalences.  But yes, there is a very significant portion of BoM cost which is invariant across generations. 


XG-PON is indeed 2.4 Gb/s up, 10 Gb/s down.  ~4:1 rate asymetry seems to be acceptable at least for consumer services (although the Freepers might not agree with that). 


More to the point, cost of moving to 2.4G upstream is pretty reasonable, while 10G upstream cost would be extremely challenging.  The good news is that the bandplan would place it in a slightly negative dispersion region of the fiber, so direct modulation might still be possible (at least without reach extension).  The bad news is that a cooled package would be required, with attendant manufacturing cost and power drain.  On the OLT side, there is a very significant engineering challenge in building burst receivers with adequate sensitivity and dynamic range, especially if guard time is kept short (consistent with GPON and BPON).


You're absolutely right that metro and regional networks will have to evolve along with the access.  I assume that there's some pretty sophisticated modeling behind the engineering rules they're using to dimension those networks.  As service rates increase (regardless of the rate of the access), these networks will have to be reinforced.  

paolo.franzoi 12/5/2012 | 4:25:48 PM
re: Moto Finishes Two XG PON Trials

Duh,


Sophisticated?  Maybe. 


I don't know but I got the feeling that at least Verizon never got the idea that the PON was its own traffic management domain.  I suspect a couple of the real techies I worked with did, but much of the networking folks have always assumed that access is the equivalent of a leased line.


Other than that, I concur 100% with your thoughts.  People really overlook the power bits of things.  For example, I took a long hard look at Flywheel Battery Technology once.  Boy that could really do some things for us all if we could shrink it some and get the manufacturing easier.  No memory effect.  No heat issues.  Very long life.  Sigh....


seven


 

HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE