The nice thing about standards is that there are so many to chose from. Engineering.com article Moving Towards Product Innovation Platforms put a great deal of focus on opportunity to use standards to solve the problem of data integration and interoperability in product development, engineering and manufacturing.
I captured few very interesting quotes. Here is the one by Gartner’s Marc Halpern:
Halpern said, “I do believe that standards should be playing a greater role. It’s really funny how, among my manufacturing clients, they all want standards and they all want interoperability but they don’t seem to make it a top priority when they are choosing software partners.” He added that, “If there had been standards supported by the vendors in question, then that wouldn’t have been as much of a problem.”
Another one comes as a “dream request” from Prof. Martin Eigner:
The bottom line, Eigner claims, is that what is needed is to define and create better international standards is to bring all the different systems together. “The dream is that we would have a high-level product structure on top of all the existing systems and that we also had a high-level architecture of change and configuration processes. Maybe we can do it on top of the legacy system.”
But the following passage from Volvo Group’s External Collaboration Manager Patrick Langton. is my favorite, because it explains the nature of the problem and key complexity – existing data and systems.
“We have parts of the hub already in production. We call it step one, or WISE 1, where we support the initial phase of collaboration. It’s in production and we’re planning to develop it into the next phase by implementing our service architecture. But that is the tricky part right now, and it’s not a hub that solves that. It’s really our legacy system that is the challenge right now.” Langton added that in the end, Volvo Trucks need to move to some kind of standardization, especially for the definition of a BOM. “That’s right, it’s so different. Some companies are very manufacturing oriented, some are very product development oriented. So, we need to find that definition. If Volvo can define it and convince our partners that we have a good definition, that will be the standard.”
The idea of magical standards that can help to organize disparate tools and data into platform isn’t new. Some of these standards are really useful. JT and STEP were specifically mentioned by Daimler Mercedes’ former IT director, Dr Alfred Katzenbach, but at the same time he is asserting that a real question is how to combine JT and STEP to make the best of both.
These comments made me think about standards and data integration problems. The main challenge PLM environments are facing is heterogeneity between enterprise systems and product development domains. Each product under the umbrella of PLM innovation platform has its own schema and the way it manage data. If we think about universal plug-n-play architecture two things needs to be handled – schema map and match. It will help to translate data for queries and transactions in the universal way.
The underlined reason for heterogeneity is the fact schemes are created by different people and different reasons. Even two programs written by two people for the same reason will use different data schema, variables and names. Also, each PLM implementation is a bit different. Although PLM systems and databases are created for the same purpose, PLM systems and implementations contain a lot of variability.
With all these challenges, to reconcile two databases or even data sets produced by different systems in manufacturing company can be vert challenging. Here are some typical problems you can face during the integration process:
- Semantic is not fully captured by data models of applications or even standards.
- The keys in schema and data models can be unreliable .
- Semantics of information is subjective and can differ even for the same data
- Combine data correctly is difficult and it can require lot of handwork.
Despite all these levels of complexity, community and analysts are pushing for standards. Although standards is something most of customers like, I can see two main reasons why even if the best standard will be created tomorrow, it will not solve a problem in practice.
The first is related to existing data and schema. Usually company already has something implementation. In the example of Volvo Trucks above, the legacy system is one of the challenges. To make a change and apply a standard is very costly. Company usually has not enough incentives to make a change.
Second one is related to domain delineation. Think about data, systems and standards. There are some overlaps. Where is the end of engineering and beginning of manufacturing? Where is the boarder between PLM and ERP, MES and PLM, etc. It creates layers that can be hardly reconciled and standardized for specific company.
What is my conclusion? Standards is not a silver bullet to create PLM innovation platform. In practice, standards can work for only limited use cases where the number of attributes is relatively small and there is a strong incentive to agree on a standard (or exchanging data is critical for business process). I can see some places in engineering and manufacturing processes where standards can work. CAD file translation is probably one of them. Procurement could be another one. Standards can be good a reference point for companies to make PLM implementation. It can remove some level of data integration complexity. However, when it comes to complex product development processes, BoM management, change processes, etc., implementations will always run into some level of diversity and ambiguity. Because of that, I don’t see how standards can be used to establish a universal plug-n-play PLM innovation platform. It is probably a question for PLM technologists and architects to find a new technology to solve challenges of product data integration complexity. Just my thoughts…
Disclaimer: I’m co-founder and CEO of OpenBOM developing a digital network-based platform that manages product data and connects manufacturers, construction companies, and their supply chain networks. My opinion can be unintentionally biased.
Pingback: Beyond PLM (Product Lifecycle Management) Blog » Intelligent part number is dead. Long live meaningful attributes?()