Data Analytics To Drive IC Shift

The adoption of predictive analytics has the potential to drive the next round of IC industry innovation and growth. Much of the necessary data handling technology is now available from other sectors. However, to fully capitalize on the possibilities, the IC manufacturing world faces particular challenges in figuring out how to get a high yield of actionable information from its streams of varied data, and how to selectively share subsets of it securely with suppliers and customers. SEMI introduces the new Smart Manufacturing Forum at SEMICON West 2016 on July 14 to put it into perspective.

“The next five years will see many more sensors everywhere in the fab,” suggests Ben Eynon, senior director, Engineering Development, Samsung, Austin, who will outline future fab trends at the SEMICON West’s Smart Manufacturing Forum. “Fabs are all still really struggling with abrupt machine failure, and we could mitigate the problem a lot with more sensor data. We’re already working down the list on existing machines on what takes down the machine hard, and in the near future we will need a whole lot more sensors everywhere, tracking vibration on rotating spindles, variations in temperature in more places, particulate levels in chemicals, optical sensing of accurate positioning.”

Next, new disruptive approaches will be needed ─ to store, merge, manage and analyze these new massive amounts of data of all different structured and unstructured types fast enough to be useful. “Perhaps we need an approach that floats a processor over the data instead of piping it all over to the processor,” he suggests. “The closer to real time we can do the analysis, the more valuable it is.”

Finally, there is the issue of security vs. collaboration. Fabs need the tool makers to be able to find and fix problems as quickly as possible to get production back up, and tool makers need some access to the process data to be able to get production running. Solutions for sharing limited, masked data, such as perhaps only the changes in parameters, or only the snapshots at a particular time, will likely evolve, to convey the needed information without revealing anyone’s secret sauce. “We have to figure out ways to share the kinds of data we need to share, while still protecting core IP on each side,” notes Eynon.

Paradigm shift will require collaboration through the supply chain

The IC industry is on the verge of a paradigm shift, driven in part by the emergence of IOT-based sensor and big-data enabled technologies,” suggests Thomas Sonderman, VP and GM of Software at Rudolph Technologies, who will discuss these issues in the program. “Traditionally Moore’s Law paced the rate of innovation, and wafer-size transitions drove the evolution of operations technology,” he notes. “But now the best place for innovation to enable the next cycle of advancement is the rapid adoption of smart manufacturing techniques ─ specifically the use of data for proactive decision-making. By establishing a secure environment where relevant data can be used across the supply chain, the entire electronics ecosystem will benefit. However, it’s only going to happen if the industry establishes a mindset built on trust and collaboration.”

As the supply chain becomes increasingly interconnected, not all the data a company needs to best serve its customers is within its own walls anymore. Chip designers need enhanced process capability information from their foundries to optimize their designs, equipment makers need detailed wafer quality data from their tools to improve performance, and system providers need visibility through their supply chain to better predict and control delivery of products. “To get actionable knowledge out of all this data, we must adopt secure mechanisms for access and transmission,” contends Sonderman. “A core requirement for the next generation of IC manufacturing will be the ability to automatically sanitize, distill, and analyze digitized signatures from across the supply chain.”

Application of these technologies in advanced packaging may particularly benefit both suppliers and their customers. “The complexities associated with the various flavors of fan-out and 3D stacks on costly, known-good die currently require a lot more experimentation than people are typically comfortable with,” says Sonderman, noting that defining intelligent methods for fabless companies to collaborate with OSATs and equipment OEM’s will be essential to quickly monetize the results from these investments. “Connecting the dots, requires accessing the dots,” he points out. “Innovation will require that companies across the ecosystem be able to trust their data to be accessed and utilized in a mutually secure and profitable manner.”

Automated, user-specific tagging could help secure IP
One approach to the knotty issues of sharing the data and context needed to improve results without fear of compromising IP could be an automated secure gate keeper such as offered by NextNine. This enterprise-level system to manage the security of the manufacturers’ operational technology in SCADA/ICS environments automatically discovers assets, and then runs all access to them through a secure central point. Using encryption technology developed for the military, it identifies each computer or equipment node by checking a random selection of the node’s multiple unique identifiers, and then allows that node access to only particular pre-set levels of information. “NextNine allows a single point of entry for all access to the enterprise, and controls each person’s access to only the specified information allowed to them,” says Don Harroll, NextNine U.S. Sales Manager, who will discuss IP protection issues at the SEMICON West forum. The company claims over 6,000 industrial users for its software, which is also sold by five of the seven major industrial automation suppliers under their own brands.

SEMI members have formed a Smart Manufacturing Advisory Council and a special interest group to start laying the groundwork for the necessary collaborative efforts and share information. The group will look at the various communication protocols and standards, models for IP protection and encryption, the current state and desired future state of the various levels of the manufacturing automation flow, and other consortia and other industry’s best practices. “We need to figure out how to hand off the information from SECS-GEM or other standard formats at the base semiconductor level to the next layer of printed circuit board assembly and systems makers,” notes Tom Salmon, VP of Collaborative Technology Programs at SEMI. “The front-end semiconductor world has the legacy of being ahead of the curve with its M2M (machine-to-machine) information and control. We’re good at it, and the packaging, assembly and EMS worlds are starting to look at taking advantage of the updated interoperability standards.” Long-term, sector players hope to extend the electronics sector’s best known methods to other automated industries, instead of having less suitable standards pushed down upon semiconductor manufacturing.

Other experts invited to discuss these issues in the Smart Manufacturing Forum at SEMICON West on July 14 include Amkor SVP Package Products Robert Lanzone, Applied Materials VP New Markets & Services Chris Moran, Global Foundries VP IoT Rajeev Rajan, Intel VP IoT/GM Industrial Anthony Neal Graves, Microsoft Product Line Manager Big Data Analytics Bill Jacobs, Optimal+ VP WW Marketing David Park, Qualcomm SVP Engineering Michael Campbell, and Siemens VP Business Development, Electronics, Sia Langrudi. Learn more about the keynotes, executive programs, eight forums, and more at www.semiconwest.org.