Remember the metro-area Ethernet hype of a few years ago? Like most great ideas, it was based on a simple notion: Since nearly all data begins and ends its life in Ethernet frames, why not run Ethernet across the metro network rather than using services based on ATM and Sonet?
Corporate IT understands Ethernet. And, IT was told, services based on metro Ethernet would be half the cost, provision faster and allow bandwidth to be "dialed up" virtually instantaneously and in more granular increments than Sonet-based services could offer.
The idea of metro Ethernet resonated with corporate IT but struck fear into the hearts of the incumbent local exchange carriers (ILEC), who preferred to squeeze every last dollar out of their multibillion-dollar ATM/Sonet infrastructure investment. And who could blame them? After all, some 70 percent of their revenues comes from voice traffic, which Sonet excels at transporting efficiently and reliably, says Burton Group Corp. analyst Dave Passmore. But their future growth depends on servicing the growing need for data networking services as well.
While the ILECs considered ways out of this dilemma, including putting a friendly Ethernet interface in front of their Sonet infrastructures, start-ups such as Yipes and OnFiber Communications jumped in and began offering native metro Ethernet services. They delivered as advertised, and the pressure was on.
That is, until the telecom bubble burst. Then the ILECs retrenched. And most of the start-ups -- after burning through hundreds of millions of dollars and overextending themselves in a mad race to build metro network infrastructures -- suddenly went into bankruptcy. The ILECs breathed a sigh of relief, and metro Ethernet faded from the headlines.
IT professionals gritted their teeth and went back to provisioning those T1s. A few brave souls began lighting their own fiber for some interoffice links, but the cost of last-mile interconnects and the complexity of building and managing a multipoint fiber network over a metro area left all but the bravest IT organizations on the sidelines.
Now metro Ethernet is re-emerging, along with a new 10 Gigabit Ethernet standard that offers unprecedented scalability. Yipes is emerging from bankruptcy with scaled-back coverage and a more realistic business plan. OnFiber has gobbled up weaker competitors. And the ILECs are promoting metro-area Ethernet services. But what you see may not be what you get.
Some ILECs are experimenting with metro Ethernet services using carrier-class Ethernet switches. But most are pushing "next-generation Sonet," hoping that if they can present an Ethernet interface on each end, users won't care how the data is transported in between.
Putting an Ethernet front end on Sonet makes it more transparent to end users. But when it comes to delivering on the full benefits of metro Ethernet, users may discover that the emperor has no clothes. These systems simply move the Sonet entrance ramp from the customer premise to the carrier's point of presence and require expensive retrofits to the carrier's equipment. And bandwidth allocation is still less efficient than with native Ethernet (although at least one equipment vendor is said to be working on this). So it may be easier to connect, but you'll likely pay more for these services than for a native Ethernet design.
Certainly, Sonet has its strengths. Proven reliability and very fast fail-over rates make it attractive for traditional voice traffic and mission-critical applications such as real-time transaction processing. Although vendors have developed technologies that let metro Ethernet meet this challenge, the products aren't yet in wide use. And the Metro Ethernet Forum is still working out a common set of carrier-grade service classes, management specifications and service interoperability standards. Much of that could be ironed out some time in 2003.
But for many data networking applications today, metro Ethernet would be the best alternative -- if it were widely available. If users wait until carriers have fully depreciated their Sonet infrastructures, it could be years before metro Ethernet becomes ubiquitous, analysts say.
But corporate IT organizations have the power to speed up that process by demanding native metro Ethernet services now. That means seriously considering services offered by companies such as Yipes, local cable operators that are lighting fiber and early ILEC offerings. The gamble of going with smaller providers can be mitigated by lower service costs and by the establishment of lower-bandwidth backup services from the ILECs, which should be clearly told that they've been relegated to second banana in those metro areas.
No vendor wants to leave business to its competitors. The ILECs will have an incentive to respond more quickly, lest smaller, nimbler vendors eat their lunch. Either way, in the long run, corporate IT wins.