PLM Data Architecture Evolution For Dummies

PLM Data Architecture Evolution For Dummies

This time of the year is a great opportunity to reflect on multiple topics. My article last week – PLM System Architecture Evolution for Dummies generated a great number of comments and discussions online and offline. Thank you very much- you helped me to shape my next blog to talk about the fundamental piece of PLM solution – data architecture and data management. If you missed the article above, I highly recommend you to check it first before continuing to read this blog.

Data is the key and the most fundamental part of any PLM solution. And it is a cornerstone of any PLM architecture. The importance of data architecture is explained by the need to collect, preserve and manage the data about the product, all its changes, histories and share this information with your team, organization, suppliers, and contractors. These requirements are simple and complex at the same time. Products are getting more complex as well as relationships between companies during the design, engineering, and manufacturing process. Altogether it brings the need to have a data management solution to support it.

PLM data management architectures and tools were evolving based on the available technologies and the needs of engineers and manufacturing companies. In the picture below you can see my version of PLM data architecture evolution.

I classified data management layers to help you in the understanding of the data architecture logic (1) Data Storage; (2) People and Organization; (3) Data Architecture and (4) Data Modeling. I’m going to describe how technologies were evolving and what data management technologies are available today.

1- Data Storage Systems

First PDM/PLM systems used network file storage to keep data and files. Later, a single database became an ultimate source of data storage for PLM systems. Combined with file storage (vault) it provided a place to store data. The demand for distributed solutions brought various replicated technologies that used mostly for files. Database clusters were a solution to scale data storage. Modern data storage solutions include a combination of multiple databases (polyglot persistence), cloud storage, and also more rely on data as services solutions.

2- People and Organizations

The next layer of data management solution is to provide an abstraction level for users, organizations, and other related structures. First solutions mostly oriented to working groups (teams) having access to the same files. For a very long time, databases logically provided a source of data for a specific group of people in the organization, which was followed by a company abstraction and later single and multi-tenant architectures. Multi-tenancy can be confused because it can be supported on the application servers level and on the data level. The first allows for sharing the same servers for multiple companies. Data multi-tenancy allows more granular data modeling and sharing of data between companies. The most advanced multi-tenant data management systems provide a network architecture capable of managing tenants, data, and relationships and access control across multiple tenants.

3- Data Management System

This is a core foundation of all data technologies. Early architectures used proprietary databases and solutions to manage data. But for the last almost 3 decades, SQL database was a key foundational piece of every data solution and PLM tech. Later systems added a variety of data replication technologies, multiple servers. Later on, search databases (technologies) became popular and useful to improve data accessibility. Modern data architectures rely on polyglot persistence principles and extended usage of NoSQL databases. Abstracting the data layer and hosting it via cloud providers gives more options in the data management toolsets. The architecture of data access, servers, and services was initially a single computer (sometimes the same one as to run a CAD system). Within time, databases moved to dedicated servers and added data replication technologies. Servers can run in company data centers or become hosted to support cloud deployments. Modern SaaS technologies and products have highly available architecture that relies on micro-service architecture.

4- Data Modeling

Initial data architectures used proprietary schemas to describe the data. As the system moved to SQL database adoption, schemas became predefined in the database. It was a step forward but required many changes during system deployment and implementation. Later PLM technologies invented a flexible abstraction layer to define data elements and made data model schemas flexible. It used internal architectures, sometimes called object-relation modelers (also called data model configuration, administration, manager, or similar.)  Because data was defined inside one logical database, it is difficult to exchange and share data between companies using even the same PLM system. Data federation was one of the technologies that enabled references to reference “external” data located in a different instance of the same PLM used by other companies. Federation is also the technology to link PLM systems to other data management systems. Modern data architecture has mechanisms to identify data in the context of each tenant and by doing so creates new ways to make data globally available and shareable between tenants (users, teams, companies and global organizations including their supply chain).

Data Management Semantics and Features

A separate aspect of data management and data modeling is related to specific features of data models to support various applications, customer requirements, the semantics of the applications, and application best practices. These are important topics, but it is too big for this article. I will talk about it later.

What architecture is used by PLM products in the market?

Not all vendors are the same and many aspects of data management are not shared by all vendors. Earlier PDM systems used proprietary databases and could operate from a single computer. Solidworks PDM workgroup is a good example. You can still see this product even it was discontinued by DS. Most advanced, mature PLM systems in the market are using SQL-based database technologies to run their applications capable of supporting tens of thousands of concurrent users in global enterprises. Aras, Enovia, Teamcenter, Windchill are all in this group. Earlier cloud products hosted using public or private cloud are also using SQL databases as a foundation of these technologies. I think Arena and Autodesk Fusion Lifecycle and maybe some other applications are belonging to this group. The technologies are evolving transparently from customers. Also, platforms are evolving and using additional services integrated with the existing ones. Modern SaaS platforms and products such as Autodesk Forge, Onshape, OpenBOM are using a set of databases and various modern data management technologies. Some other products are using specialized platforms (PaaS) and don’t have independent data management functions. These systems are fully relying on data management capabilities provided by such a platform. An example of such a system could be Propel PLM, which runs on top of the Salesforce.com platform.

If you have information about what data management technologies are used by each PLM product, please share. I think transparency in data management technologies is important, especially when it comes to supporting, future enhancements, and preventing data locking. The open secret is that data locking is still one of the elements of most business models in the market.

What is my conclusion?

The data management technologies changed dramatically for the last 30 years and so, the capabilities of PLM systems. Each application has its own trajectory and a lifetime. Mature technologies always look perfect. Think about gasoline internal combustion engines these days. The technologies of these engines are perfect, but the future is maybe still imperfect but electric. The same happened with phones, airplanes, and other products. PLM data management is the same – you cannot stop progress and innovation. I’m very close now to answer the question of what is the value and importance of each of these PLM data management technologies for customers and how to choose the right one today. The answer is different and depends on the specific situation for each customer group and specific customers. Stay tuned for the next article. 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 networksMy opinion can be unintentionally biased.

Images credit RRZEicons, CC BY-SA 3.0, via Wikimedia Commons

Share

Share This Post