Optical integration reaches the mainstream

OvumOptical integration’s transition from a niche to a mainstream revenue generator was highlighted in my ECOC 2012 presentation.
Demand for optically integrated products can bring significant ecosystem changes, prompting companies developing silicon photonics to enter the transceiver market and enabling OEMs such as Cisco to manufacture proprietary solutions.
Developing these products requires new tools and deep pockets, but we are encouraged to find multiple low-cost fabless ecosystems emerging. We reported on OpSIS earlier this year and were introduced to Effect Photonics at ECOC.
Although these are early days, the success for optically integrated products and vitality of optical communications depends on supporting competition, innovation, and driving low cost.
Optical integration reaches the mainstream
Optical integration was a key ECOC 2012 theme featured in tutorials, workshops, technical presentations, and market focus. My presentation argued that optical integration is becoming a mainstream optical communication revenue generator, with the expectation that over 50% of the market will use integrated components by 2017, an increase from about 25% today.
Moreover, transmission products for the line and client side will join wavelength management in generating significant revenues from integrated optical products. Line conditioning, mainly amplifiers, is the only segment where less than 50% of the products are integrated.
Line- and client-side transceivers at 40G and 100G are the main contributors to this transition to optically integrated products. Line-side growth comes from products including the nested modulator and coherent receiver. Client-side optics had very little integration prior to 40G and 100G, but the IEEE standards supporting WDM and parallelism and the market need for high faceplate densities have changed that.
Competition from component and OEM suppliers
The 100G client-side opportunity is forecast to grow to more than $1 billion per year, and new products are needed to support client demand. Low power consumption is critical to delivering high-density solutions.
Consequently, many suppliers developing products for chip-to-chip interconnect are also now looking at the 100G client market. In our recent report Silicon Photonics: a Tale of Two Markets (September 2012), we caution that the requirements for modules are different from those for short-distance interconnect (tens of cm) and they have different challenges.
Cisco’s Lightwire acquisition provides another potential attack on the traditional ecosystem. Lightwire provides unique silicon photonics based technology that can support low power consumption and high-density modules.
Cisco may adopt a proprietary transceiver strategy to lower cost, decrease time to market, and build competitive barriers. It need not go through the standards process, which would enable its competitors and provide them with its technology. Cisco only needs to convince its customers that it has a robust supply chain and that it can support its product.
Cisco is not the only OEM that might pursue this strategy. For example, Huawei acquired CIP (Center for Integrated Photonics) a hybrid integration developer, early this year. Huawei is interested in developing its own transceivers where it makes sense, and the emerging market may be such a situation.
We believe other OEMs such as Alcatel-Lucent will also seriously consider this course. As OEMs have done on the line side for coherent 100G, they are now looking for competitive advantage on the client side.
Optical integration challenges merchant component suppliers
Optical integration and OEMs developing proprietary client-side transceivers significantly change the market dynamic. Some important points include:
  • Suppliers need deep pockets to compete. OEMs are better capitalized than optical component vendors and have an advantage in developing integrated products.
  • If OEMs execute such a strategy the merchant transceiver revenue opportunity would decrease, hampering its financial stability.
  • Product success will be driven by superior performance. Suppliers need access to semiconductor manufacturing facilities and customer specification knowledge. Vertically integrated OEMs will be better informed with specification information.
The issues of packaging and testing, still among the highest of product processing costs, become even more significant as performance on these measures can serve as a differentiator.
Effect Photonics, OpSIS offering key optical integration services
Effect Photonics and OpSIS offer optical integration fabrication services to lower product development cost by sharing facilities among many and developing libraries, processes, and rules like those used in the complementary metal oxide semiconductor (CMOS) industry.
Effect Photonics is similar to OpSIS with a broader material system. It is offering a generic fab service that supports InP, SOI, and SiO2/SiN. It started as ePIXnet, as part of the European Framework Program.
ePIXnet supports research and industry in the area of photonic integrated components and circuits. Effect Photonics is developing a generic fab model that currently focuses on six basic elements and one common process.
The elements include a waveguide, amplifier, phase modulator, reflector, photo detector, and a tunable DBR (distributed Bragg reflector) mirror. It touts that generic integration allows multi-project wafer runs and better process control. Effect Photonics is attempting to apply to photonics the model that made electronics successful.
Daryl Inniss is practice leader for components telecoms at Ovum. For more information visit www.ovum.com/