One important difference is that the market for networking gear has started to come back. Also working in NPU vendors' favor is that OEMs are becoming more reluctant to risk developing an ASIC device when they can buy a standard, programmable processor that has already been taped out. In some ways, it's a theme that has become the new mantra for the chip industry as a whole.

OEMs have gotten smarter about making the transition from hardware to software methodologies, learned about the intricacies of network processor architectures and become more familiar with the development tools. In the near future, network processors may be the only way to go, proponents say.

AMCC, for one, thinks NPUs are best equipped to handle the expansion of core networks. The key is finding an NPU architecture and programming methodology "that enable system designers to treat the entire NPU function as a single processing thread that handles the start-to-finish processing of each [packet]," writes Steve Klinger, field applications engineer at AMCC.

Some of the credit for the NPU renaissance should also go to the industry-standard platform known as the Advanced Telecom Computing Architecture (AT-CA). An article by Uri Cummings, co-founder and vice president of product development at Fulcrum Microsystems, says the pairing of ATCA platforms and NPUs "can remove the limitations to design reprogrammability and reuse that have remained a promise for too long."

According to a recent report from In-Stat/MDR, favorable market conditions and technical improvements mean NPUs have a "bright future." NPU sales reached $60 billion last year and could approach $300 billion by 2005, the analysis firm predicts.

Though NPUs may not be right for everyone, the evidence suggests they have made enough progress to warrant a second look.