Can Big Data Solve the Counterfeiting Problem?

The electronics supply chain has been trying almost from its inception to prevent or weed out counterfeit components. Solutions to date include marking components with a unique identifier and procuring parts only through authorized channels. But several companies in recent months have suggested that following the data generated by a component’s manufacturing, test, QC and related processes may be a better anti-counterfeiting method.

Optimal+, a vendor to major semiconductor manufacturers, provides software solutions for collecting, cleaning and aggregating data from multiple manufacturing locations. Its systems currently focus on improving manufacturing yield, quality and productivity. Company executives envision applying the data generated by those processes to create unique identifiers for components. That data can then be used for verification purposes.

“Companies already collect data as chips are manufactured and tested,” said Optimal+ CTO Michael Schuldenfrei. “Customers should be capable of verifying their purchases against that repository.”

Every component, explains Schuldenfrei, already has identifying characteristics including batch numbers, date codes and test data. By adding an electronic chip identifier (ECID) during manufacturing or test, every part becomes unique. “With our current solutions, identity data is already complied and stored in repositories at each chip supplier,” he said. “Component buyers could then compare their devices to the data in the repository. We already have a lot of the ingredients needed to solve the [verification] problem.”Optimal-Slide2

The Optimal+ vision for component verification starts at the original component manufacturer (OCM). An electronic chip identifier (ECID) or an unclonable ID (UID) would be embedded in or fused on to a device during its testing process. The identification and other data would be stored in a repository.

After the components are assembled onto a board at an EMS, the board typically undergoes functional test at which point the test system can read identifying data from the components and verify them against the various repositories. To simplify and coordinate the process, Optimal+ is proposing a central hub through which the authentication requests will flow.

All of this takes place in real time, said Schuldenfrei. “We’d be applying [manufacturing and test] data in a way that will enable any party in the supply chain to authenticate a component at any time.”

But Schuldenfrei is also cognizant of the challenge that Optimal+ — or any solution provider—faces regarding anti-counterfeiting efforts. First, OCMs and OEMs would need to buy into a system or practice. Buyers would need a way to compare the components they bought with the devices in the repositories. Access rights to the repositories would have to be determined. All this requires a level of data-sharing the supply chain isn’t accustomed to.

Right now, said Schuldenfrei, individual companies don’t even share data across their own divisions. “In many cases, each factory operates as a silo,” he said. “Even central management doesn’t have visibility [across the manufacturing spectrum].” OEMs, he added, may not even want to see everything that goes on in their supply chain: they count on their partners to get things right.

Current anti-counterfeiting solutions have already run in to problems with industry buy-in. An initiative backed by the U.S. government’s Defense Logistics Agency (DLA) uses ink-based plant DNA to identify individual components. However, the DNA ink isn’t applied at the OCM factory—a majority of OCMs haven’t licensed the technology. Instead the DNA ink is applied farther down the supply chain. Without OCM buy-in, DNA opponents argue, there is no way to be 100 percent certain that a device hasn’t been counterfeited or swapped somewhere along the supply chain.

Analytic software provider Paradata also has a data-driven solution to counterfeiting. Paradata has developed a system that culls through data to find and fix weak links in global supply chain networks. Paradata’s predictive analytics software enables supply chain transparency, which, in turn, helps companies uncover potential problems even before they happen.

Paradata founder Scott Slinker, a serial entrepreneur, founded a fraud-protection company that analyzed data to verify a person’s identity. He figured the same concept could be applied to the supply chain. As a component moves from production to consumption it leaves a verifiable trail of information in its wake, he explained. Paradata analyzes the data for gaps, anomalies, and patterns. Paradata software can verify and authenticate every part, every person, every location and every event involved in the manufacturing of a product. The three-year old start-up plans to expand into new verticals soon including automotive and industrial.

One of the benefits of the Optimal+ or Paradata model is cost. Optimal+ hosts its data in the cloud. Paradata provides its software as a solution (SaaS). Other anti-counterfeiting solutions require a robust IT infrastructure or special scanning equipment to read component identifiers.

Optimal+ ‘s Schuldenfrei acknowledges investment may be necessary for any industry-wide solution. “There is investment in doing anything like this and gauging whether the ROI is justified. I think OEMs will have less of a problem [with the cost]. The further downstream you go you get into issues of liability and quality—that’s where I think the solution will bring tremendous value and easily justify the cost.” Both solutions are also non-invasive: some verification procedures can damage chips while they are being tested.Optimal-Slide12

It may take a big OEM such as Cisco to push the supply chain toward an industry-wide solution said David Park, Optimal+ vice president of worldwide marketing. “You could use the stick or the carrot,” he said. “A lot of it depends on the relationship between the supplier and the customer. The [Optimal+ solution] isn’t as easy as supply-line contracts. But we do believe visibility [across the supply chain] benefits the end customer.” Optimal+ already has a chip-test database of about 50 percent of the chips manufactured today and processes 35 billion devices annually.

Real-time data solutions can improve overall electronics supply chain integrity and security by connecting partners through the collection and storage of chip IDs and the “test fingerprint DNA” data from billions of chips, Optimal+ executives said. This will enable multiple authentication and identification flows that can be shared by semiconductor and electronics companies. “What we really see is the opportunity to extend the Optimal+ [semiconductor] solution deeper into the supply chain,” Schuldenfrei concluded. “We already have many of the ingredients we need.”