Zaffire, Inc. http://www.zaffire.com announced today its Z3000 DWDM platform for metropolitan and regional networks. The vendor says that it is cheap enough, dense enough and small enough to finally make the technology affordable for service providers to deploy in volume.

The company has been working hard over the last few months to redefine itself. First it was a name change, in March. Now, a slimmed down and focused product strategy (see Zaffire: New Name, New Approach). Industry watchers like what they see. "In the beginning their message was a little muddled," says Scott Clavenna, principal analyst with Pioneer Consulting http://www.pioneerconsulting.com. "But this DWDM product is really impressive in terms of port density and power consumption. And they're pricing is supposed to be at least half that of their competitors."

Zaffire is not alone in its pursuit of DWDM in the metro network. Long-haul network DWDM players like Sycamore Networks Inc. http://www.sycamorenetworks.com, Ciena Corp. http://www.ciena.com, and Nortel Networks http://www.nortel.com have all announced plans to re-develop their long haul products for the metro network (see Sycamore Comes To Town). Start-ups like ONI Systems Corp. http://www.onisystems.com, which is about to go public (see ONI's IPO Price Range Raised), and Sorrento Networks Inc. http://www.sorrento.com are also targeting this market (see Sorrento Puffs Itself Up For An IPO)

But so far DWDM in the metro network has been a flop. Why? The main reason is that it has been too expensive to justify deploying it, especially for those carriers that already own their own fiber. Often it is cheaper for them to light up another fiber than deploy a big DWDM system.

"When you start talking about building out the local network there is this initial outlay of capital," says Tim Burke VP of local services planning and development for Qwest Communications Inc. http://www.qwest.com. "If you have 'the build it and they will come' mentality then you could spend $80,000 to $100,000 on a box that may not be able to bring that much back on your investment right away. If vendors can bring that cost down to $20,000 or $30,000 and allow providers to grow incrementally, more providers would see the value."

Long-haul players have tried to solve this problem by removing expensive amplifiers and using less expensive components. But when there are only two companies, JDS Uniphase and E-TEK, supplying the market with components, variety and cost savings aren't that easy to come by (see JDSU and E-TEK: The Omens Are Good ).

"The difference between Ciena and Sycamore and some other player is not very great, " says Clavenna. "As a result there isn't much distinction in function or in price."

This is where Zaffire's approach differs from its competitors. Instead of buying components from suppliers, the vendor has designed its own and outsourced the manufacturing. Doing this has cut the cost of the Z3000 by almost two-thirds, claims the vendor.

"It makes a difference not being at the whim of the only two component vendors supplying the market," adds Clavenna.

One other advantage to this approach is that Zaffire has been able to build an inexpensive amplifier that it says will extend the reach of its product from 100 km up to 400 km. This extra distance allows providers to consolidate metro rings into a regional ring, which increases the number of customers served, and as a result increases the revenue generated with the installation of a single product.

The vendor claims that its third big advantage is that the Z3000 packs up to 256 wavelengths into a box that takes up one-sixth of a telco rack. Its small size also means that it consumes much less energy, lowering provider costs even further.

This seems to be right on target with what providers are looking for. "Anytime you reduce the footprint and power consumption of a device, I'm happy," says Wesley Ford, senior director of transmission engineering for Cable & Wireless http://www.cablewireless.com. Another important feature is support for "digital wrappers," a concept started by Lucent Technologies http://www.lucent.com that appends logic to Sonet data chunks to allow the same data to be carried on a raw optical stream. Conceptually, this means that providers can monitor capacity and correct bit errors on a per wavelength basis, so that as traffic goes from one region to another and over the long-haul network, providers can automatically provision links end-to-end.

"The idea of digital wrappers is a good one and it is significant that they are working on this," says Clavenna. "But there are four or five standards out there right now and if the one they are using doesn't work with the one the long-haul carrier is using then it won't be much help," he adds.

The validity of any vendor's claims must always be tested with real life customers. Zaffire has named BroadBand Office, Inc., http://www.bbo.com and Williams Communications Group Inc. http://www.williams.com as two that are currently testing the product. It has also alluded to the fact that Juniper has been conducting interoperability tests with the Z3000. But all these endorsements should be taken with a grain of salt. Zaffire and BroadBand Office are both funded by same VC, Kleiner Perkins Caufield & Byers http://www.kpcb.com. Williams and Juniper are both investors in the company.

And even if the Z3000 lives up to the low price point it promises, and fully utilizes the other features mentioned, Zaffire may still have a tough time selling to carriers. The problem is that as a start-up Zaffire is only offering one part of the solution and if a bigger player, such as Nortel, can offer all the pieces, that could be more attractive to a carrier looking to minimize head-aches.

"The real business case for buying from one vendor or another isn't the initial capital cost," says Burke of Qwest. "It's the cost of maintaining and managing it in the network. If you can get the add-drop mux, the DWDM and the optical switch all from the same vendor, it's a lot easier to manage and maintain."

Even so, some think that Zaffire is onto something. While the bigger players didn't see Zaffire as enough of a threat to comment on the new product announcement to Light Reading, analysts covering the market are watching very closely and with optimism.

"There needs to be a real set of conditions in the metro network in terms of traffic and product flexibility to make a compelling justification for a special product," says Tom Nolle, president of CIMI Corp. http://www.cimicorp.com "I think that Zaffire does that."

by Marguerite Reardon, senior editor, Light Reading

http://www.lightreading.com