Startup Promotes Parallel Optics
Paracer's products address the need to move tens of gigabits of data over short distances from board to board inside telecom equipment and among different pieces of gear inside a central office installation. It's making transmitter and receiver pairs using arrays of 850 nm VCSELs (vertical-cavity surface-emitting lasers) and detectors, which are connected by a fiber ribbon.
Paracer's first product is a 1x12 array with an aggregate capacity of more than 30 Gbit/s, says director of marketing and business development Naveen Bewtra. He claims it is "shipping to all major OEMs."
The other products, to be introduced next month (September 2001), will be a 4-channel transceiver (4x3.125 Gbit/s), and a transponder that's compliant with the Optical Internetworking Forum (OIF) standard for OC192 (10 Gbit/s) VSR-1 (see OIF Sets Short-Range Sonet Standard).
One reason that Paracer's gotten off to such a flying start could be that it was founded by engineers who already had a great deal of experience in making parallel optics modules. Founder and CTO, Dr. Lee Xu, was previously senior engineering manager for the optical transceiver group at Tyco Electronics. Before that he was with Hewlett-Packard Co.'s fiber optic communications division, which later became Agilent Technologies Inc. (NYSE: A). Both Agilent and Tyco were early developers of parallel optics technology.
Likewise, Dr. Heng-Ju Cheng, founder and director of electrical engineering, was previously a chief designer for various transceiver products at Tyco. Prior to that he worked for Agilent on high-speed integrated circuits, including those for parallel optics modules.
Paracer's arrival brings the number of companies in this space to about a dozen. Apart from Agilent and Tyco, these include Alvesta Inc., Corona Optical Systems Inc., Infineon Technologies AG (NYSE/Frankfurt: IFX), Picolight Inc., and Zarlink Semiconductor Inc. (NYSE/Toronto/London: MLT), formerly Mitel.
The sheer number of companies targeting parallel optics indicates that it is maturing as a technology. It's no longer enough to say "I have samples, hear me roar!" Now, companies need to do two things. First up, they need to convince customers that they can deliver the product in volume. (Early parallel optics modules were like gold dust, and only big-name customers could get hold of any.)
Bewtra says that Paracer's got this issue in hand. It has signed up Foxconn Electronics Inc. to manufacture its parallel optics modules, and a line has already been dedicated to that purpose.
The other thing vendors must do is differentiate themselves, and they're mostly doing that by trying to pick a convenient form factor for their modules. Some companies, such as Corona, are trying to automate the process of placing optics modules on a board (see Corona Claims 40-Gig First). Others, like Paracer and Picolight, figure that the industry will gravitate towards pluggable solutions.
Pluggable modules make it possible to test the electronic chips before the optics are placed on the board, says Bewtra. This approach also allows customers to repair or upgrade their optical modules without throwing out the entire board. What's more, they should also be more reliable, he contends, since the optics are not subjected to the high temperatures of soldering.
Paracer's modules are edge connectable, which means they can be slotted into the edge of the board without having to remove the board from the rack. Picolight's modules, on the other hand, are vertically pluggable, which means it's time to get your boards out, guys.
Sierra Ventures is the main investor in Paracer. The startup received $6.5 million in financing last fall (November 2000), which is enough to allow it to reach medium volume manufacturing, says Bewtra.
— Pauline Rigby, Senior Editor, Light Reading