Bluetooth will recognise other devices within 9m. There will be no exchange of data between devices, however, until a so-called pair-and-bond procedure takes place.
Pair is the discovery process in which the devices recognise each other as Bluetooth devices (sounds a bit like Sun’s Jini technology to me). Bond is the equivalent of telling your device it is okay to communicate with the other device.
First implementations will be rather unsophisticated, such as a Bluetooth-enabled headset that can pair and bond with a Bluetooth-enabled cellphone. Essentially, when you turn on your headset and mobile phone, the phone displays a message indicating that it found the headset. The user types in a pin number that comes with the headset, thereby pairing the headset from then on.
The first cellphones with Bluetooth capability will ship in Europe from Nokia and Ericsson. Users in the US can expect the same capabilities by the first quarter of 2002. Nokia is implementing Bluetooth on its 6210 by giving users a Bluetooth-enabled battery pack.
On the handheld side, companies such as Socket Communications will offer compact flash type 1 cards for any Windows CE handheld priced in the $US150 range.
In the first iteration, Bluetooth devices will be point-to-point. A handset will communicate with a printer, a digital camera or a cell-phone.
Piconet — a Bluetooth SIG (specific industry group) term — will follow Bluetooth. Piconet allows multiple devices to form an ad hoc network of as many as seven devices with one host talking to seven clients. If one client or workstation wants to talk to another, it has to go through the designated host. The host is the device that creates the frequency-hopping pattern, according to James Blackwell, a senior application engineer at Cambridge Silicon Radio, which has offices in the UK and Texas and makes a one-chip Bluetooth solution.
A client must communicate through the host to talk to another client.
The follow-up to Piconet is called Scatternet, which will allow groups of ad hoc networks of seven to communicate with other ad hoc networks, again, only through the host of each network.
The questions that remain are: Will companies buy in to and support these technologies, and couldn’t the same ad hoc networking functionality be created using a wireless LAN and some kind of network administration tool?
Ephraim Schwartz is an editor at large in InfoWorld’s news department.