Q&A With Fred McClimans
|
| Site Editor: | What's the status of the metro market today?
|
| Fred: | The metro market is actually exploding. But it's exploding in two different directions. There's a tremendous infusion of capital to build it up. And there's an explosion of bandwidth demand from the periphery of network, which is occurring in a different time space continuum than in the metro network.
|
|
| Site Editor: | So, the periphery and middle of the network initiatives are not running in the same place in time?
|
| Fred: | What we are seeing is classic East-West train disconnect. You build trains from both sides of the country and they derail or don't match up when they meet in the middle.
It's a classic commuter crunch-random, myriad roads coming into the beltway. Networks and highways are similar in many ways. Usually doesn't work as well as people expect it to. But when there's a break in the network, caused, for example, by changes in traffic, there's a pretty high price to pay. And in some ways, a massive beltway network isn't the best way for people to get to center of network and back out.
|
|
| Site Editor: | What is currently driving technology at the edge of the network?
|
| Fred: | If you look at the edge of metro network, it's connecting up to essentially five different types of networking environments. There's the large enterprise organization that right now is dominated by Ethernet technologies. There's the small-to-medium business dominated by Ethernet. And in the residential market the network within the home is dominated by Ethernet. The infrastructure of network service providers also dominated by Ethernet.
|
|
| Site Editor: | Isn't the center of the Metro Network decidedly un-Ethernet-like?
|
| Fred: | Long-haul carriers are pushing SONET. The core of the Metro network is being driven toward SONET, which is only one of the five initiatives at the edge of the network. That doesn't mean SONET is bad, but the market should rethink its strategy.
|
|
| Site Editor: | Why the growing interest in Ethernet?
|
| Fred: | It's coming about for a number of reasons. It's been facilitated by the tremendous increase of capacity of Ethernet within the network. Gigabit per second Ethernet is common as is 10 Gbps or even higher. And the general state of the Metro network demands that the industry take a hard look at Ethernet.
|
|
| Site Editor: | What are the competing technologies within the Metro network?
|
| Fred: | Well, there's traditional routing model where providers are deploying Level 3 and above switches. That's the classic Terabit switching approach. It says that you should build a bigger router at the center of network. But it doesn't address the higher performance issues or simplicity issues that need to come into play in the center of the network. Bringing complexity into the center of network is not what we need.
There's also the very popular, bigger, better SONET approach. But that's geared more toward the networking model of 10 years ago, which anticipated a more circuit-driven environment with more voice traffic in the network. So now we find ourselves in a situation where those approaches aren't necessarily the best approach.
|
|
| Site Editor: | Where does Ethernet come into the picture?
|
| Fred: | Ethernet is being proposed as an alternative. Not just extended Ethernet but Ethernet as a high-speed transport mechanism. Couple that with advances in electro-optical switches and you suddenly get a system that can meet the performance requirements at the center of the network.
|
|
| Site Editor: | But bringing Ethernet into the core isn't a new proposition. Weren't earlier implementations less than successful?
|
| Fred: | That is a key point over early Ethernet implementations in the Metro network. Extended Ethernet-or Transport LAN Services-probably didn't have the distance or performance reliability of internal Ethernet. It previously used low speed frame relay to hook up low speed 10 Mbps LANs. It didn't work too well. There's not enough performance. Others also tried extending the distance of regular Ethernet over the Metro network. That gave you pure point-to-point connections with no switching. Distance was limited to a kilometer or two. TLS went way of dinosaurs for a couple of years.
|
|
| Site Editor: | Why does Ethernet stand a better chance in the Metro network now?
|
| Fred: | The switching has improved; the distance limitations have improved. The changes in traffic are tremendous-voice and data are going across the packet network. We have a very different environment today than we had 10 years ago with the same basic desires.
Today's optical Ethernet offerings are so much more than the TLS offerings of 10 years ago. You have to think of them as sort of necessary evolutionary step to get to where we are today. Think of them as trees off the primate fork that didn't evolve into human beings.
|
|
| Site Editor: | Haven't other technologies stepped up to the plate? Don't they offer sufficient speed and performance for the Metro network?
|
| Fred: | In 10 years we've had a number of technologies trying to fill the void in center of the network. First, frame relay came along, with a high speed of one and half Megabits; then there was T3, promising 4 -5 Megs. But Ethernet is already at 10 Megabits. And though ATM came along to take networking into the 100 Megabit range. But 100 Megabit Ethernet has already come along with 10 Gigabit on the way. SONET stepped up and said can d o 1 Gig or 10 Gig but Ethernet is progressing at a rate that will outpace SONET.
|
|
| Site Editor: | Won't the marketplace forego all the progress made in the Metro network by moving away from SONET now?
|
| Fred: | Not really. A lot of the elements that Optical Ethernet uses are compatible with SONET. That's a good thing because we learned a lot of things through SONET about the Metro network. If we didn't learn from it, we would step backwards. But we have to build a technology more appropriate to the time at hand.
|
|
| Site Editor: | Which players are going to win over this marketplace?
|
| Fred: | The big guys that dominate the market today are going to dominate the market tomorrow. This marketplace is not going to make another Cisco. The reason: it's an extension of another market. Cisco and the Internet were less of an extension of an existing market and more of a departure. The chances of success in the market increase with the density of the vendor selling products in marketplace. The Nortels, Ciscos, Lucents, Optels, and Ericssons are players that know how to deliver service and support of networking on a large scale.
|
|
| Site Editor: | Is there any opportunity for other players?
|
| Fred: | Beneath that there's room for a reasonable number of emerging companies to deploy technology into this space. They have a great opportunity to not only position themselves as early providers but to become partners with, or parts of, these larger players over time. We expect there will be a good bit of acquisition and mergers over next couple years as larger players acquire some of the new companies to get low-cost solutions. And the middle tier-like Juniper, Boundary and Extreme-will move in and realize that to compete, they will need to have broader product lines. Players in the second tier need to be aggressive
|
|
| Site Editor: | What about the ELECs? Is there a market for bypass providers in the Metro network?
|
| Fred: | Absolutely. I think it's a bit premature to list ELECs as a separate entity from CLECs. But people will be looking for price per bit with management. If you can provide a good story, infrastructure, and pricing, it doesn't matter. If what you provide is what the customer is looking for, you're in good shape. So, some CLECs and ILECs will become ELECs. Those providers will probably gravitate toward this type of solution faster than other players in the marketplace will.
|
|
| Site Editor: | What are Ethernet's strengths as an infrastructure technology?
|
| Fred: | The nice thing about using Ethernet as an infrastructure is that Ethernet does scale extremely well, a heck of a lot better than SONET does. SONET data rates are pretty damn expensive. And Ethernet is cheaper-25 percent of the SONET price. And, Ethernet leverages all of the switching equipment and technology already out there. It doesn't require deployment of new technology until you reach really high speeds.
Now what we're looking at is Ethernet really being able to scale as rapidly as the switching infrastructure can grow.
|
|
| Site Editor: | Are there any caveats?
|
| Fred: | No one's going to dominate the market right now. Even if you can be first to market with product., it doesn't matter. The market won't mature until the middle of 2001. It's going to take Metro service providers and vendors to settle down and look at long-term plans. It's not more than an idea right now. They haven't figured out how to make a big chunk of revenue off of these services? There are no pricing models. Most of the Optical Ethernet providers are willing to talk about pricing models are coming in at half the price that we see from the massive players coming in with their multi-Terabit SONET.
|
Meet the Leaders of
Optical Ethernet.