Frankly, my first impression was that this company is insane. The explosion of multisource agreements is a signal that OEMs-which wield tremendous buyer power-don't want to be tied to any individual module maker. Incorporating Layer 2 functionality into a transponder goes against the grain. Even a large module vendor would have trouble pulling it off.

Furthermore, OEMs with their own Layer 2 ASICs will be reluctant to buy a module with integrated Layer 2 processing. (Network Elements also sells more traditional Layer 1 transponder modules.)

OEMs will generally prefer to use a Layer 2 product they are already comfortable with. If they have not already picked a Layer 2 chip, then they are likely to seek out a best-of-breed product among those on the market. There's a low probability that one company, any company, will have both a best-of-breed transponder and a best-of-breed protocol processor.

To date, Network Elements has received $100 million of funding. Since transponder-and protocol-processor-design and application support have very few economies of scope, it seems surprising that a startup would pursue both business models. Companies capture value in the transponder market with either superior manufacturing automation or with in-house active optical-component (laser or modulator) manufacturing expertise. Companies capture value in the protocol-processing market by getting to market early with a robust product and plenty of applications support.

Despite all the negatives, though, I was intrigued by the concept. Network Elements claims to have several design wins lined up-validation that the concept is viable. And since Layer 2 functions are largely standard, the technical barriers to pulling this off are not as large as they seem at first glance. Finally, the combination makes sense in that it can speed time-to-market for new system products.

I will watch this integrated-module business with keen interest. If the startup is successful, it will readjust the balance of power between OEMs and module makers and claim value for itself in the process.

Jeremey Donovan (jeremey.donovan@gartner.com) is a principal analyst at gartner dataquest.