Look Ma, No Moving Parts!
Just under a year ago, Lynx Photonic Networks launched a solid-state, 8x8-port optical switching subsystem (see Lynx Claims Optical Switch Advance). This switch appears to be one of the frontrunners in all-optical switching technology, since it has actually shipped to customers, according to Lynx.
Now Lynx has bettered itself. Today, it's introduced a family of fully-integrated subsystems that can add or drop channels as well as cross connecting them (see Lynx Unveils Photonic Subsystem). These will be on display at the upcoming ECOC exhibition.
According to Lynx's president Michael Leigh, the startup has been sampling 8x8 port OADMs (optical add-drop multiplexers) for the past nine months, and is now gearing up to start shipping them in production quantities before the end of the year. It also has a 16x16 port version that's already in the hands of development partners.
Lynx has four partners, the only one that it has announced publicly being Infineon Technologies AG (NYSE/Frankfurt: IFX), which is also an investor. Other investors include ADC Telecommunications Inc. (Nasdaq: ADCT) (see Lynx Closes $30M Round).
Lynx sees its OADM products being used in conjunction with core routers and switches. "OEO [optical-electronic-optical conversion] needs to be pushed out to the network edge," says Leigh. "What we are encouraging our customers to do is look at photonics for express processing and OEO for local processing."
To demonstrate the applications for its product, Lynx built a proof-of-concept OADM system comprising three nodes and 20 wavelengths, which was on display at the recent N+I show in Atlanta. "It was 18 crates of equipment, and took 5 guys a whole week to put together," recalls Leigh. The system is now back home in Israel, where its available for teleconferencing demonstrations, he adds.
At each network node, an OADM can drop channels for local processing, let others continue intact, or do both drop and continue -- a function that's often needed in vitual private networks where information is intended for multiple destinations. "We're the only fabric on the market that can do this -- the rest are simple on--off switches," Leigh claims.
It's worth pointing out that Trellis Photonics is working on a method of switching using holograms, which also supports weighted multicasting and broadcasting. But the startup has slowed down its development plans significantly this year (see Trellis: 'No Shut Down').
Inside Lynx's OADM subsystems, the architecture is broadly the same as its earlier switches. Where Lynx has advanced is in terms of complexity, rather than performance, which it figured it had pretty much optimized already. The guts of the OADM is a silica-on-silicon chip imprinted with rows of tiny 2x2 thermo-optic switches -- rather than 1x2 as in the earlier switch products -- connected by waveguides.
In terms of the number of individual switching elements on a chip, the 16x16 OADM is probably the most complex thermo-optic switching fabric that's ever been created. "It's a very similar architecture to the 8x8 subsystem, but it's not double the complexity, it's four times," Leigh notes.
Only Japan's NTT Electronics Corp. (NEL) is known to have developed anything similar, but its 16x16 port switch was just an R&D project, Leigh says, and it didn't include the control system, which is a key part of the subsystem.
The control system is the part that configures the heaters on individual switch elements to turn them on, off, or somewhere in between -- splitting the available light between two outputs in a predetermined ratio. This is what makes it possible to support functions such as drop and continue, weighted multicasting and broadcasting.
Leigh sees manufacturers of switching subsystems based on MEMS (microelectromechanical systems), such as OMM Inc., as the main competition -- mostly because their products are actually shipping. Previously, OMM has argued that Lynx's switch would occupy far more space than its MEMS subsystem - a criticism that appears to be borne out by the amount of gear that Lynx had to ship to its recent N+I demo.
"The main threat [to Lynx] is that carriers refuse to respond [to the case for all-optical switches]," says Leigh. "Clearly there is less pressure now than there was when there were 300 CLECs [competitive local exchange carriers] around."
Incidentally, Lynx has a separate development team working on a lithium niobate version of its thermo-optic switching subsystem (see Lynx Adds to Switch Suite). This is a niche technology, whose nanosecond switching speed makes it suitable for building optical packet switches.
— Pauline Rigby, Senior Editor, Light Reading