WDM-PON Standards: Playing the Long Game

Posted on March 21, 2012 by Pauline Rigby

WDM-PON might not get standardised before 2020. That possibility was raised at the 3rd WDM-PON Forum Workshop, which took place last month in Munich, Germany, alongside the FTTH Conference 2012.

“Under NG-PON2, it is most likely there will be no WDM-PON,” asserted Klaus Grobe, principal engineer at ADVA Optical Networking, who presented an update on standards and research projects.

That doesn’t mean standards work has stalled. On the contrary, the Full Service Access Network (FSAN) group is expected to publish a white paper soon on the technical options for the next-generation of PON standards, which go under the working title of NG-PON2.   FSAN’s job is to get everyone to agree upon a single proposal, which it then passes to the International Telecommunications Union (ITU) for standardisation.

Five approaches being considered are what might be termed “true” WDM-PON, in which each customer terminal receives a single dedicated wavelength. These approaches fall into two broad categories: tunable-laser-based systems, and “seeded” approaches, where a broadband light source passes through an arrayed waveguide grating (AWG) to create “seed” signals that transmitters can lock onto.

An alternative approach, tipped as the favourite, is “stacked XG-PON” – essentially four XG-PON systems operating at different wavelengths on the same fibre. XG-PON is the 10Gbps successor to GPON, so this approach would have a total capacity of 40Gbps downstream and 10Gbps upstream, shared among all subscribers on that branch of the PON.

WDM-PON proponents say that the higher bandwidth enabled by WDM-PON would be cheaper in the long term; it puts operators ahead of the bandwidth demand curve and thus reduces the number of investment cycles in new equipment. Mass deployment of WDM-PON would require a significant reduction in the cost of components, but literally dozens of companies and research projects are chipping away at this problem.

However, operators may favour a more incremental approach that would be cheaper to implement, and “stacked XG-PON” fits that description. Unlike true WDM-PON, it doesn’t require replacement of splitters in the field, and has the potential to be backwards compatible with previous generations of PON equipment.

If stacked XG-PON1 becomes the preferred approach to NG-PON2, then standardisation of WDM-PON would have to wait until the following standards cycle. Allowing five years between technology generations, this would mean WDM-PON would probably get standardised around 2020.

In the meantime, WDM-PON has been making progress on a different front: in backhaul and enterprise connections. In the week of the conference, the ITU approved G.698.3 (formerly G.sdapp), which outlines parameters for the physical layer interfaces in “seeded DWDM systems” for metro applications. The initial version of this recommendation includes seeded DWDM applications at 1.25Gbps with 100-GHz channel frequency spacing.  The work was led by LG-Ericsson and Korea Telecom, and backs up the approach that they have taken.

The enterprise market already has several WDM-PON vendors. ADVA has offered a solution based on coloured SFP transceivers since 2008. In September 2011 Transmode introduced a product that also uses pluggable transceivers, including an injection-locked self-tuning SFP. And LG-Ericsson continues to promote its Ethernet-based WDM-PON system in the enterprise market.

“If you factor in the savings in opex, then it’s fairly easy to make the business case for WDM-PON [compared to point-to-point connections] even though the capex is higher,” claimed Wim te Niet, vice president of global sales and marketing, access networks, at LG-Ericsson.

But would operators be willing to use WDM-PON in residential access networks without standardisation?

“Standards are important,” commented Jorge Bonifácio, head of network strategy, Portugal Telecom. “They are the driver for the mass market.  Nevertheless, we will deploy WDM-PON when it makes economic sense to do so.”

Portugal Telecom is looking at upgrade scenarios for GPON, and is evaluating at what point it would become cost effective to deploy WDM-PON. Although the calculations have not been finalised, the operator believes that point will come when the access network needs to support high-capacity, symmetric traffic.

Several speakers highlighted the potential for telecoms operators to merge their access and aggregation networks as a powerful driver for deploying WDM-PON. Deutsche Telekom, for example, believes it can reduce the number of central office sites to about 900, or 10% of their original number. The long reach inherent to WDM-PON (because it doesn’t use power splitters) would make it possible to widen the radius served by each central office to 50km.

These are powerful arguments for deploying WDM-PON even if the technology isn’t fully standardised.

Check out more from Pauline Rigby on Optical Reflection.

Tags
  • http://www.etisalat.ae Arif Husen

    i feel the comment, standardization not before 2020 looks to be non-realistic. Even in that case power full drivers mentioned in the end of article should eliminate the extra-cost to be invested on deploying WDM PON nodes.

    Generally speaking WDM is almost already heaviliy standardised and very minor part might be required for its deployment on the access.

    • KG

      I can’t agree (I wish I could).
      WDM-PON is pretty far away from being heavily standardized. What we have is a bit more than just a (cyclic AWG’s) grid. In terms of FSAN-like *interworking*, we have not even started. One major problem here is that there are too many *completely* different WDM-PONs which never can inter-operate. (Coherent vs. direct-detection, tunable-laser vs. seeded-reflective, remotely seeded vs. wavelength-re-use, broadband source vs. MFL, IL-FP vs. RSOA vs. REAM vs. REAM-SOA etc. etc.)
      Having this in mind, and not being over-optimistic, risk is that 2020 might not be too far away from reality.

      • ZB

        Well said. It begs a question whether and when, if at all, the industry needs to introduce a new standards. There are plenty of examples of premature, or after the fact standardization efforts. As well as succesful products without prior consensus in the standardization bodies (Infinera DTN for example). 2020? Nice round number for a press release. Why not 2019 or 2021?

  • Michael Finkenzeller

    In standardization there is a proposal called UDWDM which does work perfectly with power splitters. It basically is – as the name suggests – ultra densly packed WDM. Nice thing is it also works with AWGs.

    With respect to interworking. I think the idea is not that different WDM PON technologies (coherent, direct, seeded xyz) interwork on the physical layer. The physical layer needs to be the same e.g. coherent ;-) Within this technology the interworking has to happen.

    • Pauline Rigby

      Hi Michael, UDWDM was certainly mentioned as one of the candidates for NG-PON2, but I’ve yet to see a really clear explanation of how it works. With the requirement for coherent detection, electronic signal processing and multi-wavelength light sources (probably based on photonic integration), I got the impression that this approach was more of a technology leap than a technology evolution, and therefore its implementation was a little further out than NG-PON2. Do you think it will happen sooner?

      • Michael Finkenzeller

        Yes, it’s definitely a leap and like always with disruptive technologies a lot of explanation work is necessary. Happy to do so but my point here was that there are WDM PON technologies which work well with power splitters. To answer your question. Yes, I think it’s possible to match NG-PON2′s 2015 timeline but I don’t want to capture this blog for an UDWDM discussion.

  • http://enghusen@eim.ae Arif Husen

    Inter-operation across the vendors, I feel there are few cases where operators need to make the multi-vendor PON Access requiring multiple vendors to face each other. Usually even the case with existing standardized GPON, usual multi-vendor deployments are interdependent to each other. i.e for a specific instance of a PON network all one vendor. There are multiple vendors but they operate independently.

    That’s what some time gives an idea why to delay the potential of WDM-PON for doing non required things.

  • Stephan Rettenberger

    Coherent detection is a wonderful thing. It is the next big thing in optical transmission technology. In fact it is the key enabler to make 100G transmission systems viable for long haul networks. And it is great to see that commercial roll-out of these systems has begun – in the very core of the network.

    We need to keep in mind, however, that the cost per bit for this technology is significantly higher than for established 10G technology with direct detection. Still, carriers that suffer from capacity exhaust in their backbones are willing to pay the price penalty in order to stay ahead of the capacity demand.

    In the metro arena, however, we see a different set of requirements: operators are much more cost-sensitive and worry a lot more about space and power consumption. 100G solutions with direct detection will be the technology of choice for this segment.

    And for the access and backhaul? Well, our customers that use our WDM-PON solutions tell us that they would like to see DWDM grid and channel count, while – and this is the key point – maintaining CWDM or lower price points. Inexpensive, tunable SFPs is the innovation which will get us there, but not in the next 12 months.

    So back to the coherent UDWDM proposal: for a technology that is currently starting to see traction in the core of the network but still is way too expensive for metro networks, I predict that commercial relevance for residential access is still a decade or two away. Residential access requires sub $100.- ONUs to support $30.- per month service models. I doubt that coherent detection technology will get us there near term – and even 2020 will be optimistic.

  • Michael Finkenzeller

    Hi Stephan, of course cost is the main driver in access and this is considered in the UDWDM proposal. Coherent detection for LH/100G and Access/1G is a different animal. For example UDWDM doesn’t use Polarization Multiplexing and reduces complexity by other means. To achieve CMOS like cost reduction UDWDM will be based on Silicon Photonics. At the end of the day it’s a standardization and volume game.

  • Alexander Niepel

    Regarding the $100 ONU – well, I do not want to give a price point on a Blog, but with the perspective ADVA gave on the WDM PON Forum about a $50 tunable laser – and adding that up with a smart photonic integration, this should be rather doable! I assume the $50 have a caveeat on the volume? With new manufacturing technologies like Silicon Photonics the final price point is more determined by volume than by time.

  • http://www.linkedin.com LinkedIn

    New comment from Andre Girard, Ph.D. on the FTTx – Fiber to the x LinkedIn group: http://adva.li/tpuep

    Interesting analysis. Let me bring some additional points of interest from my personal experience in the field. Indeed I was at OFC this year and had few insight concerning WDM-PON and NG-PON.
    First WDM-PON
    It is an area that has created a lot of interest at first glance especially coming from Korea. I remember seeing them pushing a lot in Question 2 of the Study Group 15 in ITU-T of which I am a member. At that time, there were no need to standardize WDM-PON based on wavelength allocation whatever from individual OLTs or OLT using wavelength slicing from a source, whatever the source. The allocation of wavelengths was already standardized in ITU-T Recommendation G.983.3. So it didn’t fly at all. However, the Koreans came back with a new approach (seeded WDM PON) which this time created a sufficient interest to get standardized in Rec. G.698.3 this time in Question 6 (smart move). The issue here is that is that WDM-PON is still stuck with wavelength sources (more expensive), or/and use of AWG instead of splitter (more expensive), or/and the rationale for using a dedicated wavelength to each subscriber for other reasons than security. Seeded WDM is appelaing but generally limited to low bitrate and limited distances.
    The only driving reason for operators to move to anything related to PON is the economy of scale related to the total amount of subscribers supported by a single OLT. There came GPON for this reason. There should come NG-PON for that reason. The expected market for high bitrate per subscribers and WDM-PON should be the commercial market not the residential market. However, the residential market is currently fuly supported by many other solutions most of them lost cost/low price per bit. Another reason for PON is the only way for the telcos to compete against the Cable operators in North America with soemthing new that can be appealed to the subscribers the telecos lost to CATVs.

    Of course I could further discuss but time is of restriction
    Dr. Andre Girard
    Member of ITU-T SG15
    Chair of Canada IEC TC86 National Committee on Fibre Optics

    • Pauline Rigby

      New reply from Pauline Rigby on the FTTx – Fiber to the x LinkedIn group: http://adva.li/tpuep

      Dear Andre, interesting perspective. Thank you.

      • http://www.linkedin.com LinkedIn

        New reply from Saeed Anwar on the FTTx – Fiber to the x LinkedIn group: http://adva.li/tpuep

        Pauline – did you post this on WDM PON forum group site? It would be nice to hear more from the attendees on this.

        • Pauline Rigby

          New reply from Pauline Rigby on the FTTx – Fiber to the x LinkedIn group: http://adva.li/tpuep

          Well, I did post this on the WDM-PON Forum LinkedIn group, but not on their main group website. I don’t know if that’s possible… I have asked.

          • http://www.linkedin.com LinkedIn

            New reply from Bas Huiszoon on the FTTx – Fiber to the x LinkedIn group: http://adva.li/tpuep

            Hi Pauline, great recap from the workshop in Munich. I will check with the WDM PON board if we can place this on the web site of the Forum.

  • Mike Gruhlke

    Many of you might not be aware but in February 2012 the ITU ratified the ITU-T G.698.3 standard for seeded WDM-PON.

    Mike

  • http://www.linkedin.com LinkedIn

    New comment from Biplab Pal on the Optical, Ethernet and Transport Networking LinkedIn group: http://adva.li/9pshl

    Can anybody explain to me the riddle with WDM-PON? Every time I hear the cost argument. But cost of last mile to reach to each home is more than $1000. Compared to that going by SFP and mass scale DWDM fixed filter production, equipment cost per subscriber can be kept much lower than $200 or so. So what is exactly the issue? Does anybody have a report of cost analysis?

  • Charles Q. Tran

    The next 5 years is an exciting place to be and tangible future. ADVA should focus on 3 solutions: packet optical transport, carrier ethernet, and timing solution to address current market needs and demands. When we launch carrier ethernet service with CWDM few years ago with little notice. Now the market begins to see its great benefits and tremendous profit for service providers.
    WDM-PON has it own merits and benefits but the timing is not right yet. Service providers and enteprise customers are not ready. ADVA will use part of its profits to re-invest back to future. Just like any great technology, it takes time to convince market, technology to mature, and reach the inflection point. Don’t push. It took optical transport, carrier ethernet, advanced and microTCA more than 10 years of dedication to get to where we are today. Be patient and perservance.