A DISCUSSION WITH TIMOTHY DAY OF NEW FOCUS
BY CONARD HOLTON
Timothy Day is vice president, chief technology officer, and cofounder of New Focus (San Jose, CA), where he was responsible for the initial development and production of tunable lasers and now oversees the company's activities in all technology areas. Day received his bachelor's and master's degrees in physics from San Diego State University, and his Ph.D. in electrical engineering from Stanford University.
WDM Solutions: With more than 1100 employees, New Focus has come a long way from the startup photonics company I knew a decade ago. What core technologies or skills did you leverage to transform yourself into a supplier to the optical networking industry?
Day: New Focus certainly has come a long way since the beginning. The core technologies that we started the company with continue to be very important for us: high-speed optoelectronics, tunable lasers, and precision mechanics. In 1997 we leveraged these core technologies to develop tunable lasers for OEM applications and we also added precision crystal optics expertise to develop passive components for OEM applications. This transformation enabled us to develop a very strong OEM-into-networks capability, with a strong capability to integrate optics and electronics, and we continue to build upon these strengths today.
WDM Solutions: What optical component technologies are key to the next generation of optical networks?
Day: Well, I would say that tunability, high-speed optoelectronics, RF electronics, and advanced optics design remain incredibly important for next-generation networks. I would add low-cost switching and emphasize that low-cost tunable-filter technology will also enable next-generation modules. I think the importance of RF electronics and materials advances in the high-speed electronics arena such as silicon germanium and indium phosphide are also key.
WDM Solutions: Do you feel that WDM will be the predominant technology in metro, or will SONET or Ethernet prevail?
Day: I do not think that either will prevail; I think both are important.
WDM Solutions: Why?
Day: I believe that both SONET and Ethernet protocols will need access to and can take advantage of the efficiencies inherent in DWDM transport platforms. And when this happens, both will continue to be extremely important.
WDM Solutions: Tunable lasers were first seen as spares, but now are viewed as key technologies to enable service provisioning or wavelength on demand. Can they bring real economic benefits to carriers at a reasonable price?
Day: I definitely think so. There is a tremendous amount of investment going into tunable technologies and as a result customers will see solutions that are cost competitive. In the near term, transponders that enable provisioning and wavelength on demand will be able to command a price premium. This trend will enable the tunable components also to command a premium and volume and early adoption will occur. As technology investments and volume grows, costs will come down and network applications that can take advantage of wide tunability will continue to proliferate.
WDM Solutions: Which types of tunable lasers will prove most successful in which applications? Can you provide examples of metro systems now using or about to use tunable lasers?
Day: There are four basic types of tunable lasers for networking applications: distributed feedback [DFB], distributed Bragg reflectors, vertical-cavity surface-emitting lasers [VCSELs], and external-cavity diode lasers [ECDLs].
The tunable DFB is really the only technology that is currently being deployed in field applications and I believe Fujitsu is utilizing this in a metro product. Tunable DFBs have limited tuning range and are slow, but fundamentally they are very similar to the incumbent fixed-wavelength sources and entail less adoption risk.
A second type of tunable laser is based on distributed Bragg reflector technology. These lasers are similar to DFB lasers in the sense that they are also all semiconductor based. This laser can provide wider tuning than a DFB but is limited in power. As a result, it is more suitable for metro applications, but I know of no system company that has deployed them yet; however, several component companies are sampling to system providers.
A third type of tunable laser technology is based on VCSEL technology. These lasers are very early stage and have the potential to be low cost and therefore also suitable for metro applications. I know of no systems companies that have deployed this technology although Nortel has made announcements that it intends to.
Finally, the fourth category of tunable laser technology is based on external-cavity approaches. This is a proven technology that New Focus has been developing for nearly 10 years, first for scientific applications, followed by OEM applications in telecom test and measurement, and now for network applications. There are several other companies that are working on this technology as well.
Our network version of the ECDL is our fourth-generation tunable laser product and combines wide tuning range (full C-band) with high power (13 dB, 20 mW). This makes it naturally suited for long-haul applications but several metro customers are interested as well. There are no systems companies yet deploying this technology, from any component company. It is, however, actively being sampled though for use in next-generation systems.
WDM Solutions: What is your prognosis for 40-Gbit/s technology and when will it be implemented?
Day: I think 40-Gbit/s technology will be important and will be deployed but not as soon as people think. I don't think we will see 40-Gbit/s systems deployed widely until 2003 at the earliest. I believe there is quite a bit of potential for 10-Gbit/s systems to be enhanced using advanced modulation schemes and forward-error correction cababilities and this will make the benefit of 40-Gbit/s systems harder to define. This will result in slower adoption of 40-Gbit/s technology and this in turn will result in further investment and development in enhanced 10-Gbit/s systems.
WDM Solutions: You recently completed acquisition of JCA. To what extent will you become a provider of optical integrated circuits [ICs] and to what extent a user?
Day: JCA is an RF module manufacturer and as such is a consumer of ICs. New Focus has always had high-speed optoelectronics capability. For example, we have always had the capability to make our own high-speed photodetectors. We do not, however, sell these chips separately, but rather we integrate them into modules with higher functionality.
WDM Solutions: To what extent can optics and electronics truly converge?
Day: Well, this is an interesting question. I always have to ask what people mean by this. If they mean, "Will all optical and electrical functions exist on the same chip?" The answer is no. I do not believe all optical and electrical functions will be integrated into a single chip. There will always be material families, which can be optimized for different functions. For example, photodetection is best done on InP. Low voltage, high-speed (>20-Gbit/s) amplification is best done in silicon germanium or gallium arsenide. High-voltage, high-speed amplification is best done in InP. Low dispersion, wide wavelength modulation is best done in lithium niobate. I do feel that all of these technologies will continue to be important but for different applications.
I do believe you will see higher levels of integration of optical and electrical functions. For example, modulators and RF drive electronics will merge into single modules but inside them will be different materials technology. In this sense, companies with strong optics background will need to hook up with companies having strong electronics capabilities. Fortunately New Focus has developed and continues to build core competencies in both.
WDM Solutions: Do you plan to move up the food chain in terms of manufacturing more-complex modules and subsystems?
Day: We already are. Our tunable-laser modules combine advanced optics, electronics, and precision mechanics. Our RF modules incorporate multiple RF functions into one package. The same is true for our optoelectronic products under development in Wisconsin. We believe advanced components enable next-generation modules and we will continue to provide these critical components. However, as our customers and our business drive us to higher levels of integration, with advanced component content, we will support them selectively, as we have been doing. We are not interested in simply combining commodity components into modules with higher function.
WDM Solutions: What emerging technologies do you regard as critical to the success of optical networking? Are there emerging disruptive technologies that might compete or complement WDM such as optical-code-division multiple-access [OCDMA] modulation?
Day: I do believe that WDM will be here for quite some time and that advanced modulation schemes will be complementary to it. And certainly we believe that tunability will be a core enabling technology that will revolutionize the efficiency and dynamic functionality of next-generation optical networks.