All the FPGA action these days is in the new, emerging markets right? As we’ve all discussed for awhile, CPLDs are a nice steady market with slow growth, and networking apps (the ones that built the FPGA industry to where it is today) took a nose-dive a few years back and are out of favor.
According to the world’s largest programmable logic company, this line of reasoning is a bit short-sighted. While Xilinx agrees that there are new and exciting emerging markets out there, they are pointing out, with two announcements this week, that some of the new and exciting opportunities are built right on top of the old, boring ones.
For example, the world is filled with circuit boards that still use old, near obsolete discrete logic components. For the companies that manufacture products that use these boards, there is an enormous opportunity for saving cost, power, and board complexity by consolidating groups of these assorted components into CPLDs. The problem is that the logic designers that created these masterpieces of old are not exactly hanging around waiting for the opportunity to cost-reduce their work from 10 years ago. The people who actually care the most about the problem, the purchasing managers, are generally ill-equipped to make a case for taking action.
Xilinx is here to give that purchasing manager a little mutually-beneficial helping hand. Their Logic Consolidator introduced this week (press release) is a “convenient tool for quickly and accurately estimating the number of discrete logic devices that can be integrated into a single CPLD.” Design engineers beware! The purchasing department just got new super-powers. No longer can you weasel out of looking at that old design because the folks on the third floor just don’t know any better. When they come down and tell you “Hey, George, we’ve just noticed that we can replace 52 devices in our digital hyper-ventilator board with a single XC9500XL CPLD,” you can’t just brush them off with the “nah – incompatible supply voltages” argument again. They’ve got Xilinx on their side, and this time it’s even worse. They hold the dreaded bean-counter’s weapon of choice. The Logic Consolidator works in an Excel spreadsheet.
While designers are running for their professional lives, they might try escaping temporarily to SUPERCOMM in Chicago where they’ll see Xilinx demonstrating 10Gbps performance across an off-the-shelf AdvancedTCA (ATCA) backplane (press release). This proves a point beyond the usual haze of marketing-megahertz and dog-gates. You can really do 10Gbps yourself using a few handy items you probably have around the house (assuming your household includes an ATCA backplane, and one of the new Xilinx ATCA line-card development platforms.)
Here again, Xilinx returns to “dance with the one that brought them” as ATCA has the potential to be a huge market with its pernicious price/performance ratios and widespread applicability. The performance of Xilinx’s ATCA implementation will make high-end telecom designers reconsider proprietary backplane design, as the design time and complexity using the standard will likely be dramatically less for comparable performance.
According to Xilinx’s announcement the ATCA market has the potential to grow to over $20 billion by the end of 2007. Xilinx is well poised to take advantage of that growth. With their Virtex II Pro X devices (now you know what those embedded PowerPCs are for) and their recently introduced line card development boards, they have a huge lead in offering fast time-to-market for ATCA designers with minimal risk and overhead. Once those designs speed their way to market and go into production, they’ll be eating up large quantities of Xilinx FPGAs, and everyone will be smiling. Everyone, perhaps, except those among us that mistakenly thought that the FPGA telecom market was passé…