SAN JOSE, Calif. – Freescale Semiconductor is rolling out a family of three security coprocessors, targeting a market dominated by rival Cavium Networks. Freesacale claims its C29x chips, ranging in price from $99 to nearly $200, will deliver 50 to 250 percent more operations per dollar than Cavium’s Nitrox 3.
Despite its plans to significantly undercut Nitrox 3 prices, Freescale and analysts watching the company say it should still reap comfortable margins. “We think [Cavium has] been getting crazy margins,” said Derek Phillips, a senior product manager at Freescale.
The C29x is optimized to handle 2048-bit RSA encryption, the level recommended currently for secure transactions by the National Institute of Standards and Technology. It also process Diffie Hellman, and Elliptical Curve Cryptography algorithms and bulk encryption at rates of 8 to 10 Gbits/second.
The high-end chip uses three processing engines to deliver 32,000 2,048-bit RSA transactions/second at 10W. Dual- and single-engine versions hit 18,000 and 8,000 transactions/s, respectively. That compares to a few hundred transactions/s for the embedded security blocks on current QorIQ SoCs, said Phillips.
The chips require no external memory for RSA operations. However, external memory is needed to support secure boot and tamper detection features.
The first devices will ship before April. They run on a x4 PCI Express Gen 2 bus.
The C29x promises “significantly lower costs at a time when network security has never been more important,” said Joseph Byrne, senior analyst with the Linley Group (Mountain View, Calif.). “The security coprocessor space is a potentially high-margin market for Freescale,” he said in a Freescale press statement.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.