Comms chips

Comm Chips Flip Out

Hang on to your hats: The volatile communications chip industry's about to turn on its head yet again. And that could mean big opportunities for those who see what's ahead.

That's the premise behind "Comm Chips Reboot," the latest report from the Optical Oracle, Light Reading's monthly subscription research service. Despite ongoing weak demand in the telecom sector, the report states, there is renewed interest in certain types of communications chips. A new market is shaping up, and some old ideas (and their vendor proponents) are being left in the dust, as startups and invigorated older companies take their place.

Central to this new market are off-the-shelf chips that compete with internally developed ones. In contrast with traditional application-specific integrated circuits (ASICs), these chips, dubbed application-specific standard products (ASSPs), aren't designed in-house from scratch but instead are cobbled together from off-the-shelf components. These components are supplied by semiconductor vendors, but still designed for particular applications.

The trend toward these off-the-shelf ASSPs is accelerating, thanks to growing demand for them from network equipment makers, who can save costs by not having to develop their own silicon. As a result, over 60 percent of framer chips are now ASSPs.

There are three areas where off-the-shelf ASSPs are especially hot. The report provides competitive analysis of all leading products in each of the following categories, comparing key features and functions in detailed charts:

  • Next-generation Sonet/SDH chips: The report says growth is on the horizon for chips that incorporate new technologies that make the most of existing network capacity. These include generic framing procedure (GFP), which allows any protocol to be mapped onto Sonet or SDH, as well as techniques for virtual concatenation and bandwidth on demand.

  • Network processors: A new generation of faster, 10-Gbit/s chips geared to routers and application-aware switches is emerging. Network processors are a key trend because they can perform a wide range of different applications, depending on how they're programmed. This segment is particularly fertile for startups.

  • Switch fabrics: These chips include TDM (time division multiplexed) fabrics that groom Sonet channels into groups, as well as intelligent fabrics that read packet headers in order to determine which port to send the packet to.

Surprisingly, none of the emerging hot areas includes 40-Gbit/s for its own sake. The lack of spending from carriers has pushed out mass deployments of 40-Gbit/s systems, and some of the latest estimates suggest that it will be 2005 before the optical transport market recovers.

Not everyone's going with the flow as outlined above. Indeed, the emerging comm chip trends are likely to accelerate a shakeout, as some vendors are left behind, unable to achieve the financial fitness to partake in the new market.

The report groups the following 12 public companies according to the strength of their financials and their product potential:

Some surprises emerge from the report's analysis of various companies and their financial ratios. Some major players, such as AMCC and Vitesse, are showing signs of being on the verge of major financial duress. In contrast, Altera, Cypress, Broadcom, PMC-Sierra, and Xilinx have shown more stable financials and have some interesting new product potential.

Infineon and Intel haven't exactly wowed Wall Street, but both are large, diversified manufacturers with significant resources. As such, they are in a position to acquire distressed assets at dimes on the dollar, the report states.

The report also highlights a range of startups in the comm chip space, which may find the playing field widening as older companies fade out. Included in the competitive analyses are Bay Microsystems Inc., Cognigine Corp., EZchip Technologies, Fast-Chip Inc., Galazar Networks Inc., Internet Machines Corp., Mindspeed Technologies, Sandburst Corp., SandCraft Inc., Silicon Access Networks Inc., Velio Communications Inc., West Bay Semiconductor Inc., and Xelerated AB.

— Mary Jander, Senior Editor, Light Reading

Editor's Note: Light Reading is not affiliated with Oracle Corporation.
xinant 12/4/2012 | 9:54:05 PM
re: Comm Chips Flip Out It must be kidding saying that 10G NPU is a sweet
spot. This is contrary to this report:


Look how many NPU startups had dired or are diring. It is hard to believe that 10G NPU still
has a big market.
SiVlyGuy 12/4/2012 | 9:54:03 PM
re: Comm Chips Flip Out Here's a big clue for your research team: You might want to include the #1 supplier of NP and Switch Fabric ASSPs in your silly report.

ps - ASICs still rule!
RoutedWorld 12/4/2012 | 9:54:00 PM
re: Comm Chips Flip Out
ASICs still rule?

Here's a big clue for you, get out of the 90's. Do you know how many ASIC designers don't have jobs right now? Nothing rules in this market except cutting costs and making money.

Cutting costs == Cutting ASIC programs where you can buy a solution off the shelf.

ps - I have nothing against ASICS or NPUs. Both are needed in the real world.
green_beret 12/4/2012 | 9:53:59 PM
re: Comm Chips Flip Out I see lower cost FPGAs making a major dent going forward in certain applications. Any comments?
eyesright 12/4/2012 | 9:53:56 PM
re: Comm Chips Flip Out Won't work well for large scale applications. Good for smaller systems and/or proof on concepts until you go ASICs for scale.
pump up the volume 12/4/2012 | 9:53:51 PM
re: Comm Chips Flip Out Re:
Won't work well for large scale applications. Good for smaller systems and/or proof on concepts until you go ASICs for scale.

Remember we're talking telecom here - what's high volume? In commercial markets it might be 100k+. You're lucky to sell 10k of anything in telecom except for modems which makes the ASIC cost hard to justify.
Sign In