But to truly advance BTS design, it will be necessary to drive the internal bus standards within the BTS. This will motivate the silicon-level investment that's required to drive down capital costs (bill of materials) and operational costs (power consumption and reliability), and really accelerate the deployment of 3G networks.

A typical 3G BTS consists of four principal subsystems: wireline transport, baseband processing, linear transceiver and common control.

In the 3G BTS the key interfaces that require standardization are the wireline transport-to-baseband-processing interface, and the baseband-processing-to-linear-transceiver interface. To drive investment in the silicon required to lower costs and accelerate the deployment of 3G BTSes, it is necessary to standardize the wireline-fabric and radio-fabric interfaces. Without such standardization, it is difficult for silicon vendors to justify investing $10 million to $20 million, or more, in each standard product targeting this market. Similarly, individual system houses will find it prohibitive to develop their own ASICs and refresh them frequently to exploit the cost reductions that are possible via Moore's Law.

Through standardization, we should see the appearance of network processors, general- and special-purpose DSPs, linear transceiver waveform processors and control processors that natively support the standardized BTS internal interfaces. With such commoditization, 3G wireless infrastructure deployment will accelerate. For this reason, PMC-Sierra supports the Open IP Basestation Architecture initiative recently announced by cell phone maker Nokia.

Kevin Huscroft is Vice President of R&D and Chief Technology Officer at PMC-Sierra (Burnaby, British Columbia).