FTTH Technology Fracas Continues

During 2007, the world's telecom carriers connected another 9 million homes to optical fiber; by 2010, if current trends continue, the number of homes connected to FTTH will pass 50 million. A massive transition to all-fiber access is now all but inevitable everywhere. Yet the road to FTTH is full of twists and turns that make it difficult and frustrating for vendors to negotiate – not least in determining what technologies will be used, where, and when.

For a start, despite the strenuous efforts of GPON cheerleaders such as Alcatel-Lucent (NYSE: ALU), there's no chance that GPON is going to dominate global FTTH deployment any time soon. In fact, as we reveal in Heavy Reading's new report, FTTH Worldwide Technology Update & Market Forecast, the deployed bases of the three main FTTH technologies – GPON, GEPON, and active or point-to-point (P2P) Ethernet – are likely to be more or less equal, by which time around 90 million homes will be connected to fiber.

Given the hyperbole that has swirled around GPON over the past two years, this might seem a surprising conclusion. But the fact is, there is no "global" FTTH market today, and there will be no global FTTH technology – at least, not in the foreseeable future.

Why so? First of all, no two national markets are the same – and even within a single city, one technology may not fit all needs. In Japan, by far the world's biggest FTTH market with more than 11 million connected homes, GEPON is universally used to connect single dwellings, while active Ethernet dominates in multi-dwelling units (MDUs), and there's little sign that a shift to GPON in either case is imminent.

Given the size of the Japanese market, that alone means that GPON cannot win the race in terms of sheer numbers any time soon. And although GPON looks certain to be the technology of choice for major incumbent telcos in both the U.S. and Europe, active Ethernet is actually more widely deployed in Europe today – and is likely to maintain that position right through 2012.

The truth is that although the International Telecommunication Union (ITU) Full Service Access Network (FSAN) group has worked very hard to establish GPON as the worldwide reference for FTTH, its strengths – high downstream speeds, carrier-class features, and low opex, among other things – haven't convinced everyone that it is the right technology everywhere, at least not yet. For now, GEPON gear is cheaper and more widely deployed – persuasive qualities in its Asian homeland.

Meanwhile, active Ethernet advocates point to their technology's simplicity, low-cost CPE, and suitability for use in so-called "open access" networks, such as those being built by municipalities and utilites worldwide. Advocates such as Cisco Systems Inc. (Nasdaq: CSCO) argue that incumbents that deploy GPON would in effect be making it more difficult for regulators to ensure that the new fiber networks can be used on equal terms by other service providers – an argument that certainly carries weight in regions such as Europe, where DSL unbundling has driven a vibrantly competitive market.

There is plenty of room for argument around the edges. For example, which way will China jump? China has already brought fiber into or very close to apartment blocks serving at least 15 million homes – far more than anywhere else. Almost all of it is simple Ethernet switching with very little PON deployed to date, but the big Chinese telcos are all preparing to move to PON. While GEPON currently has the upper hand, the longer the Chinese telcos delay, the more likely they may opt for GPON – potentially tipping the scales decisively in its favor.

Widespread deployment of GPON elsewhere would certainly help to make the case in China, since mass deployment will drive down the cost of equipment, and cost is the big issue there: Telcos are seeking a combined price for ONT and OLT of $200 per customer – less than half the typical current price.

However, widespread deployment of GPON probably isn't coming in 2008. While most GPON vendors are bullish about the medium term, it's been a long, slow haul, with many projects running late. Verizon Communications Inc. (NYSE: VZ) has slipped behind its original schedule for GPON; Orange (NYSE: FTE), the biggest European telco committed to GPON, is also behind schedule; and vendors themselves are mostly cautious about the coming 12 months, betting that scale deployment won't really begin before 2009.

It's a Catch-22 that can only be resolved by time – and based on current trends, within a few years, carriers and vendors will need to be thinking about the next generation of technology. In addition to higher-speed P2P Ethernet, WDM PON and 10G PON are both being prepared for launch around the 2012 time frame.

Yet for all the current frustrations, this is a market manifestly worth pursuing and investing in. The transition to FTTH is a once-in-a-generation infrastructure overhaul, and the winners over the next five years will be very well placed to reap continuing benefits for years to come. FTTH is a 20-year project that will entail hundreds of billions of dollars of telco spending, with the aim of achieving domination of wireline networks for decades to come. By any reckoning, that's a prize well worth pursuing – despite the twists and turns in the road ahead.

– Graham Finnie, Chief Analyst, Heavy Reading

Page 1 / 9   >   >>
CatalystBP 12/5/2012 | 3:47:48 PM
re: FTTH Technology Fracas Continues I agree with everything you said. Having developed FTTx technology for the past 8 years, it is dissappointing that the largest markets lie outside the USA and from the looks of it will continue for some time. Unfortunately, I think there are a number of dynamics that drive this.

First, the technology is still evolving. The cost per subscriber connection, end to end triple play is still above $300 even for GEPON. This is due largely to the cost of the two principle components - optics and the interface chip, which lacks the level of integration to support triple play. Both could be solved by current technology, but the volume isn't there to warrant the investment.

Second, the system/network/management for triple play is still evolving. You need to go no further than your own blogs to see the debates over IPTV distribution, VOIP, SIP, etc. to see this is still in flux. Lot's of progress has been made here but I think it is still too tenuous for incumbent carrier deployment/operational strategies.

Third, I'm not sure the incumbents have determined how to make money off of this and protect their capex investment. There are no apps available today that would compel a subscriber to pay more for higher BW. Mostly because true broadband is not ubiquitous, so there is no market for a product that would require say 40Mb/s let alone 100Mbs or greater. (Talk about your "catch 22s") I don't believe that most users really care how many Mbs they actually get (except in Japan where its a status symbol) as long as the app performs well. So, maybe its time for a different pricing model. The FTTx technologies certainly support performance based SLAs.

Lastly, the widespread deployment successes outside the US are largely due to government mandates or subsidies. This is certainly true in Japan and other parts of Asia. As much as I despise government intervention to free enterprise, tax incentives or deferrments for companies willing to invest in access infrastructure would go along ways toward resolving the "catch 22" above. The resultant increase in real volume would drive solutions to the first two and we would be on our way to the next generation of advanced communication.

Its frustrating too see the FTTx technologies languish in this country while we watch other countries pass us by. I hope some "catalyst" happens soon to break the stalemate.
zman 12/5/2012 | 3:47:46 PM
re: FTTH Technology Fracas Continues I also agree with this article.

A few bits: Alcatel-Lucent is the technical editor of the 10-GEPON effort. They are hedging their bets. China is deploying EPON/Ethernet; they have standardized this. One other factor not included in the study is MSO increase in use of Ethernet transport in the access network.
materialgirl 12/5/2012 | 3:47:44 PM
re: FTTH Technology Fracas Continues Dear CatalystBP:
Super post. You describe the financial problem in a nutshell. Reality is that service providers do not want to be a "dumb bit pipe", but that is all they can be. They simply cannot compete with a globe of wild-eyed 20 somethings trying every new thing at the edges. So, we all sit here and try to fool ourselves, while not much constructive happens.

As to optical integration at the edge to lower costs, does INFN have anything that could be applied to that area?
jepovic 12/5/2012 | 3:47:43 PM
re: FTTH Technology Fracas Continues Agree, good article and comments.

In reference to the PON vs P2P discussion, and the issue of cost reductions, one can note that the PON volumes in general very small compared with switched Ethernet.

Also, why would China start with PON in the first place?
^Eagle^ 12/5/2012 | 3:47:43 PM
re: FTTH Technology Fracas Continues Material,

Infinera probably does not have the tech to drive cost and scale at the edge / access. They do not really do it all monolithically. They monolithically integrate a lot of lasers, EA modulators and various detector chips for monitoring. then they co-package 2 of those arrays with AWG based optical mux/demux. The PIC is too expensive for the edge / access. This is partly because the part that is monolithically integrated is based on InP technology, not silicon.

If you want to see who might be able to drive cost down optically at the end user terminal (edge / access) I would watch the silicon photonics development. Note Intel did not sell the research group doing silicon photonics work to Emcore as part of the Intel sale of the optical division.

I would watch the Intel research as well as watch companies like Luxtera and some of the research out of Europe and some universities (UCSB comes to mind as well as Berkeley, Stanford, CalTech, MIT).

The silicon photonics approach is the only thing long term that can do what we want. and only if they are finally able to get the process to work on CMOS process lines so they can take advantage of scale of silicon industry. Even if there are a few small bits of InP chips flip chip bonded to make the gain or drive. (hopefully the silicon researchers can eventually solve this without the InP chips).

So far, the most advanced work that has been talked about or published is at Luxtera and Intel as well as some work at UCSB. I am sure there are other labs working this problem.

mr network 12/5/2012 | 3:47:42 PM
re: FTTH Technology Fracas Continues Its interesting to read this article and see that the main problem with GPON is not discussed. Cost and BW.

First its to expensive CPE cost in GPON also for a foreseeable future.

Secondly it does not have the BW and in reality its only a 20-50 mbps technology which means its not suited for IPTV applications which is the killer app that drives the FTTH deployment, look at what happened to Verizon that had to start with 1:16 split and look in Japan where NTT has 1:4 or 1:8 split NOT the advertised 1:64 or 1:128. This has a tremendous impact on cost.

Thirdly it does not scale which means people are trying to introduce WDM technology overlay but the operational cost for that is very high and it does not work in reality.

So to summarize: P2P Ethernet in Gig speed is 1/3 in cost compared to 20-50 mbps GPON - this is what Europe has understood and thats why P2P will be much bigger than GPON. GPON is only hyped by Verizon since it fits their ATM legacy but it does not provide the application sto its customers and Verizon will realize this soon and be in deep shit.

Also its the small fact that every PC you buy today has a GE port inside which means that the tranceiver cost is zero compared to GPON.
jepovic 12/5/2012 | 3:47:42 PM
re: FTTH Technology Fracas Continues Edit: PON volumes are small compared with total switched Ethernet volumes (office applications, server applications etc etc)
paolo.franzoi 12/5/2012 | 3:47:41 PM
re: FTTH Technology Fracas Continues
Wrong and Wrong.

Cost - EPON ONTs go for sub $100. GPON ONTs are similar in price.

BW - PONs are inherently broadcast media - so only one copy of any given program needs to be placed on the PON. PON is very efficient for IPTV - much more so than any point to point.

Remember the 50 meg is DEDICATED BW. The uplinks will be much more greatly oversubscribed and the access link will not be your bottle neck.

Your GigE port in your PC is not 10KM single mode fiber - so you do need a transceiver.

^Eagle^ 12/5/2012 | 3:47:40 PM
re: FTTH Technology Fracas Continues Brookseven,

You stated in your post that EPON ONT goes for less than $100 and GPON is priced similarly.

is that the price for the entire ONT? (optics, electronics, housing, pcb, connectors, etc.?) or only for the optical subassembly?

jepovic 12/5/2012 | 3:47:40 PM
re: FTTH Technology Fracas Continues Wrong, wrong...

"BW - PONs are inherently broadcast media - so only one copy of any given program needs to be placed on the PON. PON is very efficient for IPTV - much more so than any point to point."

For broadcast media, as traditional TV, yes, PON is efficient (although P2P can achieve the same efficiency with IP Multicasting). For video on demand, PON is pretty much useless. Which means it is just about impossible to compete with Youtube-type of content, and you have removed one of the key selling points vs tradtional cable TV.

"Your GigE port in your PC is not 10KM single mode fiber - so you do need a transceiver."

That depends on how you build. Most active Ethernet buildouts are done with a "basement switch" in an MDU, with Cat 6 cabling to the apartments. In this way, only the switch needs a transceiver. An important point here, if one compares the switch with PON, is that the switch will give you packet-level multiplexing.

PON is a wet dream for the vendors: lots of special equipment needed, and the limited capacity guarantees a steady cycle of upgrades. Active Ethernet, in contrast, is much less interesting since the prices are far lower. The volumes of switches for FTTH are miniscule compared with the normal Ethernet switch market. Thus, there is far less buzz around this market.
Page 1 / 9   >   >>
Sign In