The Last Word
Chipmakers Remain Wary of Data Sharing
Samsung Austin Semiconductor
The semiconductor industry is in a dilemma, well described by Ben Eynon, senior director of engineering development at Samsung Austin Semiconductor, at the Advanced Process Control (APC) conference held in Austin in mid-October.
Scaling to 10nm and beyond, Eynon said, means that chipmakers and their equipment and materials vendors “must attack everything simultaneously. The problems we have are large and must be solved perfectly and quickly in order to stay competitive.”
Eynon told the APC audience “we have to keep our equipment clean and our processes right on target to keep these defects down…whoever can keep the defects down, wins. We face process limitations and integration complexity. And it takes a lot of cross-communication to make things work well.”
All of the like tools in a fleet must be matched, and stay matched. The time required for feedback and feed-forward must be minimized. And standard operating procedures in the 10nm era and beyond need to change to custom operating procedures where each wafer is treated as an individual with different processing requirements.
Chipmakers need “more sensors on everything so we can monitor more than we do today. We need feedback to adjust the next process, based on the just-prior process. I need to know that if I run this stuff this certain way, I can be sure of what I will get for electrical test results and yield,” Eynon said.
He then shifted to the crux of the problem. With so much data being pulled off the tools that already it is “spilling onto the floor,” Eynon acknowledged that the industry faces a general reluctance to share that data and lacks the necessary protocols to do so securely.
“The tool vendors have their log files. We have our process. Our customers have their designs. Everyone has their own IP that they don’t want to share. We have to figure out how to handle IP better than we do today [to make sharing happen],” he said.
INTEL’S BREAD AND BUTTER
Steve Chadwick, senior engineer at Intel’s manufacturing IT operation, described a similar conundrum. After outlining several ways that Intel engineers use proprietary software to analyze some five billion data points each day, Chadwick was asked a question about sharing data in the cloud. He replied that Intel “has its own compute farms, both distributed and centralized,” and there is no problem taking expense reports, for example, out to the cloud. But Chadwick also said Intel would be very cautious about taking its manufacturing IP to the cloud. “You are talking about a company’s bread and butter, our R&D, and there is going to be a lot more study before we do that.”
Eynon added that Samsung also uses proprietary software to improve security, including its own email system. “That’s our mentality in semiconductor. I believe we are going to be using our own hardware to do our number crunching for a long while. Every time we hear about another case of identity theft, we realize this kind of activity is not going away. We are not going to swing to an attitude of ‘here is our IP, go crunch it.’ ”
James Moyne, a professor at the University of Michigan and semiconductor industry consultant on advanced factory automation, noted that all industries face security issues, but the chip industry is particularly concerned that IP not leak. “Compared to other industries, we are really skittish,” he said. While chipmakers are increasingly sharing data with their tool vendors, only a fraction of that data is being sent outside for remote analysis.
Nick Ward, director of marketing for the services group at Applied Materials, said that “Applied Materials completely errs on the side of caution” when dealing with the security of customer data by keeping tight controls and safeguards in place.
“We are dealing with proprietary protocols and proprietary approaches to data. As an industry, we need to develop a set of standards for protecting the shared data. Being able to integrate with certain systems and tools will speed time to problem resolution and create greater cost savings for manufacturers,” Ward said.
However, Moyne cautioned that companies need to keep working on the problem, or lose out. “The tighter the control specs and the more sophisticated the algorithm, the more suppliers and customers need to share their data. That’s just a fact.”
David Lammers is an Austin-based technology journalist.