Technology Roadmapping: Observations from InterPACK

iNEMI's Director of Roadmapping, Dr. Francis Mullany, reports on an iNEMI Roadmap workshop session held at ASME's InterPACK conference and a panel he participated in discussing the roadmapping of semiconductor packaging.

Google Maps or Street Atlas?

by Dr. Francis Mullany, iNEMI Director of Roadmapping

Today, technology roadmapping is a booming industry. I was a member of a panel discussing the roadmapping of semiconductor packaging at ASME's InterPACK (International Technical Conference and Exhibition on Packaging and Integration of Electronic and Photonic Microsystems) in October in California. Four different roadmap efforts in the space of advanced packaging and heterogenous integration were represented on the panel. Unsurprisingly, the first question from the floor was, how do you differ from each other?

Technical scope is one obvious differentiator: iNEMI's roadmap extends across the full range of electronics manufacturing technologies, from materials to board assembly, whereas roadmaps like HIR focus on a range of technologies that extend up to the level of the packaged device, with a special interest in heterogenous integration. 

Another differentiator has to do with realization. All cross-industry roadmaps articulate a vision of what our industry is trying to achieve and the application and technical drivers that will get us there. Technical needs are enumerated, gaps identified and the underlying challenges are explained. What differs between roadmaps is how much detail and definition is given for anticipated solutions.

International roadmaps with a very wide range of contributors tend to limit this part to broad outlines of possible directions for technical innovation. Nationally/regionally funded roadmaps tend to be more prescriptive, outlining concrete recommendations for implementation in a particular country or economic area. It is the difference between a map with multiple possible paths to reach the destination and a specific route selection.

Of the roadmaps represented on the InterPACK panel, the iNEMI Roadmap and HIR fall into the first category; NIST-funded MAPT and MRHIEP into the second.

There are valid reasons for this difference. International, broad-spectrum roadmaps must not expose confidential intellectual property and/or commercially strategic planning detail. In contrast, national/regional-block roadmaps are usually used to inform funding decisions for particular geographies, so they recommend solutions that leverage key competitive advantages for that geography. Parts of such plans might not be published openly.

The strengths and pitfalls of each approach are roughly the same as the difference between using a routing technology like Google Maps, versus a printed road map. The former gives you step-by-step instructions on what to do – providing clarity, but occasionally routes you over a road that had unforeseeable obstructions. The latter shows all the potentially useful options, but provides less guidance on what is the optimal route for the your circumstances at the time.

My Destination is Your Point of Departure

Also at InterPACK, iNEMI CEO Shekhar Chandrashekhar and I ran a productive workshop on the topic of drivers for advanced packaging. Through a series of exercises and discussions, the participants ranked the key technology and market drivers for this technology space over the next 10 years. The top four drivers identified were:
  1. High-performance computing for AI/machine learning (ML)
  2. Photonics integration
  3. Autonomous driving
  4. Wearable electronics
The top two could be classified as technology drivers. The latter two are new application spaces with challenging environments and requirements. Interestingly, the workshop participants categorized high-performance computing for AI/ML as an application, even though it was clear that it is the emergence of new, specialized cores for machine learning and the associated heterogenous integration is what was on their minds. New specialized ML cores are an application/market driver for packaging, but a technology driver for end uses that need vastly increase intelligence.

Sounds complex? It is. We have a multi-layered map where solutions at one level (e.g., new compute cores) require mature technologies at deeper levels (e.g., packaging) meeting the needs of those higher levels, so that they in turn can deliver new end-use applications (e.g., intelligent systems).  

Clearly, extensive cross-industry technology roadmapping is an absolute must for the electronics manufacturing eco-system to avoid misalignments in the value chain and efficiently meet our industry's current and future needs.

For More Information

keywords: roadmap, electronicspackaging, electronicsmanufacturing