2013 Annual Technology Forecast

Nov. 1, 2012
From a business perspective, 2012 could be described as the year technology developers waited for a market recovery that never came.

Yup, it’s that time again. Here’s what we think will be hot in optical technology over the next 12 months.

From a business perspective, 2012 could be described as the year technology developers waited for a market recovery that never came. But the “macroeconomic uncertainty” that became the year’s most repeated catchphrase failed to dampen discussion of which directions technology development should pursue.

The first wave of 100-Gbps systems established itself in the field this year – and the second wave is nearly ready to join it. Consensus began to build on how to construct colorless, directionless, and contentionless (CDC) reconfigurable optical add/drop multiplexers (ROADMs), while “flexible grid” capabilities for such platforms approached mandatory status as well.

Meanwhile, the non-optical aspects of optical communications also took a starring role. Some of these advances involved electronics – particularly work on coherent-related semiconductors – but several were software-based. Indeed, many of the conversations this year between carriers and technology developers revolved around control planes and provisioning systems.

We’ll likely see trends that played out in 2012 continue to drive technology development in 2013. Fortunately for those of us charged with writing technology forecast articles, next year won’t be a complete repeat of this one, however. In the following pages, Lightwave will review what should be truly new and what will be “new again” in the following areas:

  1. Networking
  2. Fiber to the home
  3. Cable-operator applications
  4. Test and measurement
  5. Equipment design

As has become traditional, our prognostications combine basic research; conversations with analysts, developers, and network implementers; and a rather beat-up Ouija Board with suspect reception. We’ll also occasionally stray from what we think will happen to what we think should happen. For example, we think carriers should buy more optical communications gear next year – but to say for sure whether they will is beyond even Lightwave’s powers of foresight.

Optical networks get smarter

YES, FIBER-OPTIC NETWORKS are getting faster and more flexible. But much of the optical-network technology efforts in 2013 will focus on making intelligent use of these attributes.

You’ll hear more talk next year about making networks “applications aware.” That translates into the ability to set the right bandwidth, resiliency, and priority levels for individual services. Enabling such capabilities involves coordination of multiple layers on the OSI stack. Thus, you can expect to hear more about software-defined networks (SDNs) and whether optical communications should embrace OpenFlow or chart its own direction.

Regardless of the SDN approach, getting the control planes of equipment operating at different layers – particularly the optical and IP layers – to communicate with each other will prove essential. Several systems houses already have suggested how they’d accomplish this goal; these vary depending on where in the network such interaction takes place and whether the vendor thinks IP/MPLS, Optical Transport Network (OTN)/GMPLS, or native Ethernet processes should star. The result is a hodgepodge of largely proprietary schemes. Expect to hear more calls for standardization in this area in 2013.

More speed, please

Meanwhile, 100 Gbps will become more ubiquitous, with the focus shifting to the metro/regional. Many large vendors established positions this year on whether coherent transmission or direct detect is the way to go. Next year will see the vendors who specialize in metro/regional transport make their calls. Our sense is that many feel coherent transmission is the end game, but some may offer direct detect technology as a transitional step.

Finally, debates will continue over the best way to create CDC ROADMs. I left ECOC thinking a consensus had been established. However, a systems-vendor executive later told me he’s not so sure.

FTTx expansion should continue

LAST YEAR’S FORECAST for FTTx, specifically FTTH, was divided into good news and bad news. This year’s good news is that at least two of last year’s bad news items – fiber shortages and delays in broadband stimulus programs – have cleared up.

In 2013 we’ll continue to see focus on multiple-dwelling units (MDUs), with some of these advances affecting single-family applications as well. The most visible – well, more accurately, less visible – advances will continue in the cabling arena, where we’ll to see further steps taken to make installations down hallways and into apartments less obtrusive and unsightly. The ONTs/ONUs for these applications by definition are designed for indoor use. These units will also show up more frequently in single-family applications and will continue their current evolution toward becoming full residential gateways with integrated wireless-router capabilities.

Meanwhile, the impact of the ITU-T’s G.hn home wiring specification may finally be measurable in 2013 – which may not prove the same as being significant.

Out in the field, work on next generation PON technology also will progress. The 2012 event that may have had the most impact on 2013 technology development was FSAN’s adoption of a TDM and WDM approach to WDM-PON (TWDM-PON). The proposed architecture stacks independent (and potentially pre-existing) PONs on multiple wavelengths, with the further ability to overlay conventional WDM capabilities on the same fiber to support high bandwidth requirements such as business services. Technology decision makers at Verizon have publicly stated their preference for TWDM-PON over 10G-GPON (XG-PON1) – and sources within the PON-vendor community suggest the largest GPON deployer in the U.S. isn’t alone in this sentiment. With XG-PON1 specifications still not entirely wrapped up, we wonder how many vendors will switch 10G-GPON development resources to TWDM-PON.

Coax fends off the fiber challenge

CABLE OPERATORS and the technology developers that serve them spent much of 2012 dreaming up ways to ensure that coax networks will enjoy continued relevance in a high bandwidth world. For example, the big technological news out of last month’s SCTE Cable-Tec Expo in Orlando, FL, was the launch of DOCSIS 3.1 standards efforts. DOCSIS 3.1 will support 10 Gbps downstream and 1 Gbps upstream on hybrid fiber/coax (HFC). SCTE expects the specification to be more or less in place by the middle of next year, which means technology development will be well underway by the year’s end.

While this specification work is underway, a concurrent, if perhaps slower moving, IEEE standards effort also threatens the opportunities for optical technology in the networks of cable MSOs. The EPON over coax (EPoC) effort, as its name implies, seeks to bring the benefits of passive optical networks to cable operators without having to move off their HFC networks.

Counterbalancing these fiber retardant efforts, DOCSIS Provisioning of EPON (DPoE) appears finally ready for deployment. Version 2.0 of the specification, completed last spring, rounded out DPoE’s Carrier Ethernet capabilities. The next piece of the puzzle, the Demarc Automatic Configuration (DAC) specification that will spell out how to configure ONUs in a manner similar to DOCSIS cable modems, should be in place by the end of this year. And CableLabs has established a certification program for products designed to Version 1 of the specifications.

All of that should lead to DPoE deployments at some point in 2013, sources tell Lightwave. Most believe cable operators will likely use the technology to support business services first. But at least some EPON vendors insist that DPoE will see use in residential applications as well. Perhaps that’s true – DOCSIS-friendly EPON OLTs could be used to drive EPoC networks.

Taming the price of test

HIGH-SPEED TRANSMISSION applications continued to dominate R&D spending at many technology developers in 2012, and test equipment manufacturers reacted accordingly. Instruments for coherent transmission test and measurement applications provided the most salient example, with oscilloscope purveyors playing “Can You Top This?” with each new bandwidth claim.

It would appear that technology developers have access to test equipment with enough bandwidth to meet their foreseeable needs, sources at one test vendor recently told Lightwave. Now what they want is price reduction.

Such lower costs not only would be nice for R&D applications; they will be essential for coherent technology manufacturing and potential field test requirements. In particular, 2013 should be the year someone from the coherent-analyzer community unveils a unit designed for field use.

Back in the lab, this year saw several test vendors offer 32G bit-error-rate test (BERT) capabilities, especially for emerging 4×25-Gbps requirements. With 400 Gbps looking like the next step in Ethernet’s evolution, thoughts have already begun to turn to what the necessary electrical interfaces would need to be to support such a data rate. A paper at ECOC suggested 16×25G for the first generation, with 8×50G offered for a CFP2 equivalent and 4×100G following. So it would seem innovation in BERT for parallel applications will have to continue.

In the field, it will be interesting to see how well carriers take to Alcatel-Lucent’s announcement of integrated optical time domain reflectometry capabilities in its GPON equipment. Sources have already indicated that systems houses have integrated measurement capabilities for turn-up of 100-Gbps wavelengths. With test capabilities now appearing in PON systems, we have to wonder how many types of network systems will appear with integrated test capabilities that obviate the need for standalone instruments.

Equipment design pulls it together

IN THE NETWORKING SECTION of this forecast, we said networks are getting faster, are more flexible, and will need to get smarter. The design of systems, subsystems, and components will have to reflect these realities.

As far as faster is concerned, the building blocks for coherent 100-Gbps transmission in the long-haul are in place. As we mentioned, the focus will now turn to the metro/ regional space. For companies with in-house ASIC expertise, the ready opportunity to tweak their systems for metro use puts them in good shape. Those without this type of internal capability will have to rely on technology developed for long-haul, or they’ll have to wait for help from companies like MultiPhy, which has announced plans for a metro-focused coherent ASSP, or the module vendors (the fact that Acacia Communications has written an article on the subject in this issue is likely not a coincidence). Alternatively, they could jump to direct detect approaches. Oclaro and Finisar have displayed 4×25G transceivers and say samples of such modules have reached the hands of customers. Meanwhile, other vendors, such as Menara Networks, promise to have a similar offering in the near future.

This conundrum underscores why some observers wonder whether systems houses will start looking to do more in-house technology development at the component/subsystem level, with Cisco’s acquisitions of CoreOptics and Lightwire pointing in that direction. Certainly for the current and near future generations of coherent transmission systems, it seems unlikely that companies like Alcatel-Lucent, Ciena, and Cisco will move quickly to module-based designs until they’ve squeezed all they can from their ASIC investments.

In terms of flexibility, flexible-grid ROADM capabilities to support future 400-Gbps transmission will become more ubiquitous, as Finisar’s display of such a ROADM line card at ECOC illustrated. While no company has all the necessary building blocks in-house right now to be able to ship a fully self-sourced line-card-level product, the major component/subsystems vendors should fill in their blanks fairly quickly in 2013.

Also as mentioned previously, CDC capabilities alongside flexible grid continue to be a subject of development. Yet, uncertainty remains regarding whether a contentionless capability is really necessary. Subsystem vendors such as Finisar, JDSU, and Oclaro appear to be charging full speed ahead to support a full CDC feature set, but a fair number of their potential customers remain undecided about their interest.

The move toward smarter networks implies smarter system building blocks. For example, the flexible-grid ROADM line cards need not only wavelength-selective switches that can move off the ITU grid, but also equally flexible optical-channel monitors that can synchronize with them, either directly or through some intermediary. Line cards that can support multiple coherent modulation formats for different applications or line rates must be able to be configured and reconfigured quickly and efficiently. And as the application aware network comes into vogue, connections must be capable of being put up and torn down, or reformatted, with equal agility. In short, software-based reprogrammability will become an even more common attribute of systems and subsystems in 2013.

Of course, integration will remain a major theme in equipment design, from prepackaged line cards such as the flexible ROADM blades just discussed at the subsystems level to photonic integrated circuits and silicon photonics. There are a few startups focused on one or both of these component fields, including Aurrion and Skorpios, from whom we’ll likely hear more in 2013.