What is Missed To Connect Digital Thread?

Last month I attended CIMdata virtual event 2020 PDT and PLM Roadmap. I'm still slowly skimming the slides and connecting the dots with the notes I took during the event. You can check the agenda of the event here. The topic of the event was one of my favorites and - Digital Thread. According to the name suggested by CIMdata, It is the PLM path to deliver innovation and efficiency.

The keynote by Peter Bilello was setting a stage for the discussion. Here is the brief summary as it is presented on the agenda:

The concepts of the digital twin and digital thread have been around for decades. Industry insiders generally agree that keeping them joined maximizes their value and justifies the effort of keeping them current. Yet the gaps between them seem not to be closing. The specifics of digital twins and digital threads differ widely, but many of their underlying considerations are similar; both are fed by many of the same information flows, especially if PLM enables them. Fundamentally, a digital twin without a digital thread is an orphan, disconnected from the decisions and processes that impact it. Whether the product is a drill bit or an aircraft, its virtual representation will undoubtedly struggle to be complete and up to date without a digital thread. When information flows between them are synchronized and unimpeded, every business unit gains. Ultimately, the benefit of joining the digital twin to its digital thread is a competitive advantage. This presentation will make a case for how and why the digital twin and its digital thread are inseparable.

The key enabler of both Digital Thread and Digital Twin is the flow of the information. In my view, Peter's presentation highlighted one of the biggest problems in the industry and a gigantic barrier towards achieving Digital Thread - connections between islands of information. But let me step back and show you a few slides I captured.

The foundation of the information in the product life cycle is various representations of X-BOM.

Mindshare PLM Leaders are all presenting fascinating pictures of connected trees in a variety of forms and ways.

The dream about what will happen if…. we only are able to have accurate information available. That was the big PLM dream for the last few decades.

Unfortunately, it is easy to say, but hard to make it happen. Companies and organizations are separated by a common BOM structure residing in multiple databases and systems.

CIMdata presented a conceptual picture describing what actually should be connected. Take a look at this image.

Here is what I think is missed on this image - companies. While the image did great work presenting domains and processes, the real challenge is the notion of the companies. The data can technically flow, but the data flow is very fragmented and each company can potentially use a different PLM system and database(s).

What does it mean in practice? It means that the notion of Digital Thread is a dream of how to connect multiple databases. All mindshare PLM leaders have platforms built 20-30+ years ago with a single-tenant data model foundation. Which made me think about what is missed to make all these elements of Digital Thread connected?

Industry hopes for standards as a life-saving event. Here is a set of existing standards "in the game" of Digital Thread connecting.

While many of these standards are great achievements of companies and people groups, as Jos Voskuil mentioned in his article, there is always another standard to develop.

The benefit of these standards is also they increase the longevity of product data as the information is stored in an application-independent format. As long as the standard does not change (fast), storing data even internally in these neutral formats can save upgrade or maintenance costs. However, I think you all know the joke below.

In the same article, Jos Voskuil highlighted that OEMs have real difficulties to connect multiple companies using the same PLM infrastructure, which is another confirmation of the technological weakness of all PLM mindshare vendors.

A supplier had to download their information or upload their designs combined with additional metadata. These portals were completely bespoke and required on both sides “backbreaking” manual work to create, deliver, and validate the required exchange packages. The OEMs were driving the exchange process. More or less, by this custom approach, they made it difficult for suppliers to have their own PLM-environment. The downside of this approach was that the supplier had separate environments for each OEM.

So, where is the solution? The problem has multiple facets, but I'd like to highlight two major ones. (1) There is no infrastructure to support the digital thread in the same company and even less infrastructure to connect the digital thread between multiple companies. The last one is a real pain. Placing data in different SQL databases and separate PLM environments don't solve the problem. (2) There is no trust between companies to share information. This problem is real and hard to solve. To do so will require a combination of changes in company philosophies, business models, risk mitigation technologies, and tools.

A network-based layer capable to connect multiple companies and data pieces is needed. Each tenant in this layer must be independent, but the data must be connected. It will allow multiple companies to work together pretty much like social networks do their work or applications like Google Sheets sharing the data. Multi-tenant applications are new in the industry, but examples such as Autodesk Fusion 360, PTC/ Onshape, OpenBOM, and few others are here. Such technology is a foundation of the future connected digital thread.

What is my conclusion? A network architecture layer is needed to connect multiple players in the ocean of the information presented as a full digital thread. A completely new set of technologies is needed to solve these problems. Modern data management, which includes polyglot persistence and multi-tenant models are capable of providing a solution already today, but the adoption of these solutions might take time, especially for large OEMs and their top tier suppliers. However, building an online network by connecting pieces of information available today can be a step-by-step process to solve the problem. Just my thoughts…

Best, Oleg

Disclaimer: I’m co-founder and CEO of OpenBOM developing a digital network-based platform that manages product data and connects manufacturers and their supply chain networks. My opinion can be unintentionally biased.