Data is at the heart of every PLM technology and system. Therefore from a very early stage of PLM development software vendors put significant attention on how to create the best data modeling capabilities for their PLM system and technologies. A few months ago, I published my article where I outlined the evolution of PLM data architecture (). Please check this out. PLM data modeling started from proprietary models and now moving towards global federated data schemas with capabilities to share data with anyone online.
Data Model Flexibility
There is one thing that during the entire time of PLM data model evolution remained the most important part of every PLM data modeling architecture. I’m talking about the level of flexibility of data models. Although there is a very strong desire from almost everyone in the manufacturing world to achieve the highest level of standardization, there is very little chance that manufacturing vendors and companies will be capable of agreeing on a specific data standard or model that will define all needed data types, relationships, and semantics. Therefore, whatever data model will be coming the flexibility of the data model will play the key role and will define if such will sustain and serve manufacturing companies.
From Database Schema to Global Flexible Data Models
For many years, database technology defined the capabilities of data models. It started from proprietary databases, took some flexibility (but very bad performance) with object databases, and then expanded significantly with relational databases and the ultimate flexibility of SQL language. The majority of existing PLM platforms in production today are using SQL databases with so-called “object modelers” built on top of these platforms. These platforms are mature enough to provide reliable solutions for existing monolithic PLM systems. But these technologies are not good enough to scale in both flexibility and their ability to provide a foundation for new platforms – global, scalable, and flexible. There is a real need to build a new type of platform where data models won’t be physically connected to a specific database and, at the same time, will be reliable, flexible, and scalable. What can be a technology for such a model
Data Modeling Schemas and Semantic Technologies
What can be a language to develop a future abstract data modeling schema? There are many ideas in the industry to create such a model. Earlier, XML was suggested as a solution, later various standards in industry and PLM were suggested as an option. The solution must be abstract, powerful, and not related to a specific database. One of the most promising solutions in my view is semantic technologies and the technological elements of semantic web tech stack –RDF and OWL. These technologies are flexible, powerful, and not dependent on a specific database implementation – can be used with multiple databases as well as a fully independent tech data stack.
Polyglot Persistence Implementation
RDF and OWL can be a good foundation for the model, but how such a solution can be technically implemented, and what can be a persistent model for such implementation. If you haven’t heard about polyglot persistence, this is a time to learn about it. The idea is simple and powerful – use multiple services with different database backend implementing specific elements of data storage and semantic capabilities. In such a way, you can think about data services or entire PLM platforms exposing their data modeling capabilities in the way RDF/OWL specific data modelers can sustain and provide PLM modeling layers for data implementations. The elements of the database persistence can come from multiple vendors in a variety of licensing mechanisms including open sources and proprietary SaaS services.
What is my conclusion?
PLM industry needs to have a reliable and independent standard-oriented data modeling layer that can provide a solution to create flexible data models for the industry and not lock customers and vendors to specific vendor-oriented data models. Such a layer can bind multiple online data services and entire PLM platforms. If such a layer is found then it will unlock the global approach in flexible PLM data model standards and future development of open PLM systems. Just my thoughts…
PS. Abstract and flexible data models is a good topic to discuss for data architects in large manufacturing companies. Please contact me- I’m happy to discuss more.
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.