Optical Technology: Today and in the Future

Sprint Corp. was one of the companies that was early in its implementation of advanced optical technologies. Fred Harris helped drive many of those initiatives. Harris is director of network planning and design for Sprint, which operates numerous complex optical networks around the world. One of the most experienced network engineers in telecommunications, Harris helps guide the development of Sprint's ION (Integrated Optical Network) service, an advanced optical network for voice, data and video. Infrastructure Editor Charlotte Wolter grilled Harris on the cost and performance of optical networks.

X: What are the pressing issues for carriers, especially competitive local exchange carriers (CLECs), in the use of optical networks?

Harris: One is making the right kinds of investments in optical technology, which is changing very rapidly. The issue is finding technology to take you to where you want to go, and doing it with the right cost curves. For example, dense wavelength-division multiplexing (DWDM) provides significant capacity in the long distance arena. It has not necessarily been as attractive in the local arena. Like most new technologies, [optical] lacks maturity in a number of areas, not the least of which is in service assurance--the ability to see the performance of the customer's service in the network.

X: What are the most attractive technologies?

Harris: Optical routing is very attractive. When we get to an all-optical arena, we want to have the ability to provide survivability similar to synchronous optical network (SONET) ring technology. To get 50- to 100-millisecond restoration without SONET, a variety of technologies are being examined. They tend to use optical routing, which gets away from rings and goes to mesh [topologies], but you don't lose the attributes of SONET rings.

X: We also hear that metro DWDM equip-ment is too expensive and that very little is being sold.

Harris: Yes, but there is a lot of new technology coming into the market from companies such as Cerent (Corp., Petaluma, Calif.). These new companies are coming up with optical technologies for the last mile, and I think we will see a lot of that deployed over the next year. It is very cost-effective. The advantages of Cerent, and companies like it, is that they interface with legacy local area network (LAN) technologies, and do it with an optical backbone. ... You have the ability to provide Ethernet as if it were simply an extension of the local network. That's very attractive to many customers. Moreover, you don't run into latency problems, particularly in the metro environment, that you might have in the wide area network (WAN).

X: What other technologies could potentially have a big impact?

Harris: In last-mile technologies, there are a dozen companies that are producing or have product that will be delivered in the next six months. All have interesting approaches to solving parts of the overall puzzle of moving optical networking closer to the customer. That is going to be important because the bandwidth customers require is going to accelerate over the next few years. I think digital subscriber line (DSL) is just the tip of the iceberg. The thing to keep in mind as we do things like DSL is that application developers will find ways to use the bandwidth and then say, 'Gee, I could do a little more if I had more bandwidth,' same as with microprocessors.

X: We see companies introducing terabit routers ostensibly to feed large optical networks.

Harris: Traffic aggregation is likely to be done at the switch level and is going to move closer to the customer. Still, whatever we put out there, it has to work with what will be as well as what is, and Internet protocol (IP) has a number of difficulties working with what is. There will be terabit routers, absolutely, but other systems will be better suited to higher quality of service (QoS) and low-latency environments. The answer to that issue has yet to be determined. Our view is whatever switching fabric is out there, it will have the attributes of both IP and asynchronous transfer mode (ATM). These include the ability to control latency, guaranteed cell or packet loss, and symmetrical services--the kinds of things typically part of the traditional switched world--plus multicast, the point-to-multipoint attributes you have in IP.

X: What is the greatest challenge or need in optical networking?

Harris: The ability to do service assurance. We can keep pushing the technology envelope and get better, but when something breaks, the ability to react so the customer doesn't see degraded service is a key issue for the optical industry. They get very high marks about speed of innovation to market. It is wonderful, but the ability to do service assurance has been very slow. The technology lends itself to the perception that only when things break does it affect anything. If a laser fails, the entire wavelength goes. But you can also get degraded performance, and we must have the ability to detect that. Catastrophic failure is easily observed, but degraded performance and intermittent problems are less easily observed.

Share this article: Email, Slashdot, Digg, Del.icio.us, Yahoo!MyWeb, Windows Live Favorites, Furl
Add this article feed to: RSS, My Yahoo, Newsgator, Bloglines