The Truth About TeraBeam
Thanks to optical-editor on the terabeam thread for the following article:
The Truth About TeraBeam
Service providers that buy into TeraBeam's fiberless optical technology with hopes of living a gigabit dream could wake up to a megabit reality.
That?s the real story to emerge from a press conference held yesterday by TeraBeam Networks terabeam.com and Lucent Networks lucent.com to announce a joint venture to sell Terabeam?s ?breakthrough fiberless optical technology.?
Essentially, the conference lacked in detail what it missed in clarity. No price was announced, and neither vendor would give specifics on throughput, range or, indeed, how Terabeam?s laser-based wireless technology works.
However, some digging by Light Reading later turned up one reason why both vendors may have been unwilling to provide more granular information.
TeraBeam says its product supports ?gigabits per second speeds.? However, on closer examination it turns out that the capacity of the network is shared between subscribers -- a fact confirmed later by TeraBeam. That means that service providers that deploy services based on TeraBeam technology, and businesses that buy into those services, are more likely to see megabit speeds than gigabit throughput.
Here?s why. Terabeam?s products deliver 1-Gbit/s point-to-multipoint connectivity between a base station, which is installed on a service provider building, and multiple subscriber terminals (actually, small dishes that sit in the window of users' offices).
Despite the company?s name, the base station acts more like an arc light than a beam ? dividing each metro area into 90-degree sectors, or cells, and ?sweeping? them.
A typical 90 degree cell might contain between 15 and 40 user terminals ? up to a maximum of 96, according to Greg Amadon, chairman and CTO of TeraBeam. And the data capacity of each cell is shared between the terminals. Thus, if the cell is being used to carry gigabit Ethernet traffic, and there are 40 terminals in the cell, each customer will receive an average of 25 Mbit/s of capacity (1,000 Mbit/s, divided by 40) -- not the ?gigabit? throughput that TeraBeam and Rich McGinn, chairman and CEO of Lucent, vouchsafed in today?s conference.
This limitation of TeraBeam's architecture is worst for downstream traffic -- being sent from the base station to the subscriber terminal. According to TeraBeam's Amadon, upstream traffic, sent from subscribers to the base station, travels at "a minimum of 100 Mbit/s."
So does this mean that TeraBeam's solution is without merit? No -- for two reasons. First, multi-megabit speeds are still a huge improvement over the 1.5 Mbit/s T1 connections that many businesses are still using for Internet connectivity.
Second, TeraBeam's point-to-multipoint architecture lowers the overall cost of the network by allowing service providers to install cheaper subscriber-side equipment. Other laser-based fiberless optical products -- from companies like AstroTerra Corp., Canon Inc., FsONA Communications Corp., and Jolt Ltd. -- can only be installed in point-to-point configurations, with expensive equipment at each end of the network connection.
More serious competition for TeraBeam could come from AirFiber Inc. airfiber.com, a new startup that is due to announce its first products in the next few weeks.
--Stephen Saunders, US editor Light Reading lightreading.com |
