In the early 1990s, Intel Corp. held a significant segment of the embedded communications market with a RISC-based controller design called the i960. The company consciously threw away its market advantage when then-CEO Andy Grove decided Intel no longer would support any architecture not directly related to desktop or server.
Flash forward to ten years ago, when Freescale Semiconductor (then Motorola) held a similar enviable market share in both control-plane communication processors, in the Quicc and PowerQuicc families, and in datapath engines, in the C-Port network processor. The deterioration of Freescale's market position had little to do with conscious abandonment of the market, and more to do with the continued financial turmoil after the company's spinoff from Motorola, which had the practical effect of cutting future design engineering to the bone.
When Patrick Mannion interviewed Rich Beyer at the Freescale Technology Forum, it was somewhat understandable why Beyer focused on automotive. The communication industry stayed in the dumps for many years after the 2001 crash. In the few areas that have shown renewed OEM health in recent years, either Cisco Systems dominates the market, or new Asia-based ODM models of supply chains make new architectures difficult to justify.
Still, the conscious swing away from Freescale PowerPCs and other embedded communication architectures in certain communication market sub-sectors could be seen in reaction to Apple's decision to acquire PA Semi. Those board-level communication players who had designed in the PA architecture seemed genuinely disappointed to go back to Freescale. And the show floors at the recent Interop and NxtComm shows were dominated by the likes of ARM licensees; core specialists like Wintegra and Tensilica; and special-purpose architectures from Xelerated, Cavium, and Bay.
Freescale has a lot to prove with the upcoming QorIQ architecture. The problem is, having a single killer architecture does not help if your applications engineers are no longer close communication system designers. Freescale will have to rebuild trust and awareness from a market position that is experiencing erosion in key areas such as line cards and packet aggregators.
We should add a final observation about paying attention to the customer base. A year ago, Freescale planned their technology forum immediately after NxtComm, the largest show for infrastructure suppliers to the telecom industry. The company heard about it from angry analysts and journalists. This year, Freescale scheduled the forum directly against NxtComm. Someone is not paying attention.
