Carriers, vendors, analysts and even users like to talk about the "last mile" as a nice inside-baseball term to confound outsiders. Everybody knows the last mile isn't a literal mile; it's just a metaphor we insiders all understand, right?
But the exact location and length of the terminating loop matters -- a lot.
It's convenient for folks selling WAN "solutions" to slap the "last mile" label on anything involving the last few hundred feet, the last mile, the last five miles or the last 50 miles. Yet such labeling minimizes the importance of the real local loop by obscuring crucial distinctions between network architectures.
Many "broadband carriers," for example, are happy to promote their "last-mile" strategy as a mix of fiber and copper rentals from supposedly willing wholesale providers, even though their own networks don't pass within several hours' drive of half the population.
So you can't count on these carriers to take you out of the Bell-company local T-1 and T-3 provisioning cycle. And because spot shortages of local T-1 circuits are being reported again, that last "mile" to reach these broadband carriers may sometimes seem more like a trip to the moon than four laps around a football field.
It may also be especially important to compare terminating topologies now that DSL has been taking a bad rap. CEOs, stock prices and reputations of DSL service providers have all been dropping like flies. But DSL isn't just a service, it's a technology. It can be used in environments where weather, load coils and service-provider provisioning problems aren't a factor.
The practical application of this is in multitenant office buildings. Here the last mile isn't a mile, and it isn't even necessarily horizontal, but vertical.
Consider makers of high-speed optical-networking termination equipment sold to carriers. Some of them have led the anti-DSL vanguard, promoting fiber construction deep into office parks and neighborhoods instead. But lately they've realized that even they may have to confront old Category 3 wiring to reach the end user, making metropolitan Gigabit Ethernet and direct optical connections impractical or impossible.
That's why you're going to start seeing partnerships of optical switches with makers of in-building devices for the hot multitenant unit (MTU) market. For users looking to upgrade branch offices well beyond good old 56K frame relay, these multiprotocol partnerships could prove vital.
Why turn up your nose at these kinds of potentially powerful connections just because there's some DSL in there? DSL can be problematic when it's the last two or three miles from a telco, but fine when it's the last 400 meters in a building. Maybe there should be a new term for such a connection. How about "last lap"?
This story, "How long is the last mile?" was originally published by Network World.