New 'true optical' box targets metro core
BY MEGHAN FULLER
The buzz at SuperComm was metro, metro, metro, and one of the emerging players in that space attempted to generate some buzz of its own by staging a mock demonstration, complete with picketers whose chants and signs urged conference attendees to "Break the shackles of OEO!" and "Liberate wavelengths!"
Amid the hoopla, San Carlos, CA-based Opthos unveiled its IW 1000, an IP-managed metro core solution designed to give carriers a cost-effective alternative to traditional SONET-based and OEO architectures. The startup is touting its box as a "true optical" solution, which is not the same as being all-optical, contends CTO and foun der Joseph Parker.
"A lot of the people who claim to be all-optical still have OEO conversions some where in their fabric to do signaling and other functions," he explains. "By true optical, what we mean is that from access device to access device, we have no OEO conversions in that data path; the traffic-bearing channels are entirely optical."
The company's insistence on being labeled a true optical metro player begs an important question: Is there really a need for a true-optical box in the metro core?
"Yes," says Marian Stasney, senior analyst with the Yankee Group (Boston). "When we do OEO conversions-particularly at the level where we are trying to do transport-we add a lot of cost, in terms of delay, componentry, and complexity," she explains. "If we can limit those OEO conversions throughout the network, then we're much better off."
While the folks at Opthos remain mum on the specifics of just how the patent-pending technology works (though Stasney believes it is based on passive optics), they are obviously very eager to talk about what it can do. The IW 1000 provides what the company calls "Instantaneous Wavelengths," which enable the metro core network to change the destination of any lightpath from any port at any location in less than 50 msec.
"For the most part today, wavelengths are still treated as if they were fixed fibers," says Parker. "All they are getting is additional bandwidth. But the metro core is increasingly a dynamic place where you'd like to be able to rearrange the wavelengths to meet your load patterns, either by time of day or by application."
For example, he says, say a carrier has a storage-area network (SAN) at a collocation facility or central office that services clients within that facility. Say the carrier wants to extend that storage service to clients in other collocation facilities within the same metro region. Today, that carrier would have to lock down a separate wavelength for each of the collocations it wanted to service-which is expensive, particularly since those wavelengths are most likely not in constant use.
But if that carrier could quickly move lightpaths-say, in a few tens of milliseconds-then the SAN could control one wavelength or a few wavelengths.
In this model-the Opthos model-wavelengths become dyn amic. "When you need to have bandwidth to a particular location," says Parker, "you would request that a lightpath be set up, you would transmit your data, and you would tear the lightpath back down."
Another advantage of the Opthos box is its protocol independence; it can transport anything-from Escon and Ficon to Gigabit Ether net. "By not interfering with the protocols involved with the transport, you lose a lot of the costly overhead of packet processing," says Stasney.
The box itself is carrier-class, supporting 32 protected or 64 unprotected wavelengths in a 10U shelf that is 17.5 inches high and 12 inches deep, which is much smaller than any competing de vices, claims Parker.
The Opthos box is also unique in terms of power consumption; it uses just 100 W, while most competing products use kilowatts of power.
"When I was at SuperComm, I actually got to see the box running in their booth, and I went up to talk to them, and the CTO pulled me aside and said, 'Come here, listen to this,' and there was just silence," says Stasney. Because the IW 1000 requires so little power, cooling devices are unnecessary.
"I'm starting to see vendors rate their boxes in terms of wattage used," asserts Stasney. "You never heard about that in the past. Now, instead of 'I can give you so many ports per dollar,' it's 'I can give you so many ports per watt.'"
Stasney admits she is also impressed with the extensive interoperability tests Opthos has conducted with such access vendors as LuxN, Metro-Optix, Foundry Networks, Gotham Networks, and Nishan Systems. "[Interoperability] is so important to their customers-the service providers-today," she adds.
Opthos has chosen to work with access players for obvious reasons, says Parker: They perform the grooming and aggregation functions on the wavelengths. "Then our responsibility is to deliver those wavelengths to the correct facility within the metropolitan area," he explains.
"Our boxes interconnect the central offices and the collocation facilities around a regional area set up to about 400 km."
The Opthos box also conforms to the ITU draft standard G.692, which defines a laser grid for point-to-point systems based on 100-MHz spacing to ensure compatibility with emerging ITU-grid compatible metro-access devices.
Currently in lab trials, the IW 1000 is slated for general availability in the first quarter of next year.