PLM Architecture: Content As a Service

Picture 3I want to talk today about the future of PLM in terms of content. My initial thoughts were about SOA/WOA architecture and SaaS trends in software. I define PLM Content as a set of information related to Product Definition, Organization Processes related to product development and manufacturing as well as any valuable information that can be used for decision support on every level. What is different in PLM content is the fact that content is an information that intended for people consumption. Opposite to a lot of information we can use for application and consumption, content is a asset that provides support for people in their everyday work-life.

PLM Content As A Service. What are key business drivers for PLM content as a service? There are few, in my view:

1. Content Re-use
For me, this is probably is the most important one. PLM systems today are lack of capabilities to re-use information that already has been entered by engineers, designers, customers and many other people involved in product development and manufacturing. In many cases, we found ourselves re-developing product, parts or re-entering information such as Bill of Materials. Ability to re-use content can be a major cost reduction factor.

2. SOA Architecture
Service Oriented Architecture should become, in my view, the core architecture enables granular application organization and component-reuse. Together with REST principles, SOA is focusing on how to create flexible enterprise application architecture.

3. Reduce the cost of change
In today’s enterprise systems, change is probably the most significant challenge customers are facing. In a very dynamic business environment, manufacturing companies need to have an ability to adjust product development systems in the way that allows them to change business process, development and manufacturing practices.

What are main principles that in my view can turn PLM products to PLM Content as a service?

1. Unique creation (only one time). This is an approach to enable agile product data organization. Once it created, data can be re-used, but we need to eliminate the need to re-enter data several times. I believe, this is not what happens today. Multiple times, design information re-entered in multiple ways, such as, Bill of Materials, Drawings etc.

2. Rich meta-data. The role of meta-data is to enable an easy way to find and re-use information. The focus is to enable metadata creation and consolidation in a natural way during a product lifecycle. Rich meta-data layer will allow to build a discoverable network of product information. With such approach, PLM systems will be able to provide a single source of truth about product data and all business processes surrounding product design and manufacturing.

3. Granular user access. Future PLM systems need to allow an ability to bring only needed pieces of information for user. Simplicity is very strong way and approach in today’s computer systems. In my view, PLM still has a long way to go in order to achieve a desired level of simplicity. In order to reduce the level of complexity, systems need to be able to deliver only specific pieces of information.

4. Support for content lifecycle. Everything is changes during product development. Content services need to be able to be integrated in a simple and natural way with processes and application that manage changes of the product data. With such seamless integration, users will always be able to get right information at the right point of time.

5. Separation from application and presentation layer. This is last, but not least. Separation of content services from the user interface and content lifecycle will allow flexible application organization and optimization of user experience independently from content management. User’s demand these days is for excellence in user experience and user oriented service. New devices, new approaches are coming, and it’s very beneficial to be able to have PLM system applying these users experiences very fast.

So, what is my conclusion today? I think PLM Content As Service can be a very powerful transformation for the current PLM system architecture. PLM Content as a Service can be built around today’s SOA architecture and flexible data modeling and data management capabilities. Application of Content As A Service principle can reduce costs of PLM deployment and streamline PLM implementation in organizations.

Best, Oleg

Share

Share This Post

  • Andy Finkbeiner

    Oleg,

    Correct me if I’m wrong but aren’t most of the major PLM tools on the market place in direct conflict to these ideas? The PLM solutions that I’m familiar with are “monolithic” designs where they want to aggregate all of the user generate data into internal proprietary databases. Isn’t that the opposite of flexible and granular?

    We’re exploring SOA and PLM and the conclusions we’ve come to so far is that the major PLM tools conflict with SOA. While the sales folks might say they work in a SOA environment they really don’t because the data isn’t really available to share. So the interfaces that can be connected are superficial.

    Maybe I’m wrong on this or maybe you’re talking about a need for all of the vendors to really change their offerings at a very fundamental level? If so, that means they’ll need to throw away most of their existing code base doesn’t it?

  • Dear Andy. SOA and Content as a service are two different things. I think, all PLM vendors today already claiming SOA as supported, however I don’t see SOA is something predefined. The specific implementation of SOA principles can be very different and depends on many factors. I think, when vendors are thinking about SOA/Services they mostly exposing available API (this is might also include re-design of architecture). Content Services can be fundamental change from my standpoint. This is shouldn’t come as replacement of what companies made today with SOA, but as additional capability to expose relevant PLM content in stateless manner. It doesn’t mean companies will needre-develop existing system. And, yes, I agree, in today products data is really not ready to share. It’s possible to integrate using available API and tools, however, this process is rather complex… Best Regards, Oleg.

  • Oleg,

    I am a bit defiant towards SOA. This looks like another bloatware to me. I wrote a post on this very topic here with many references : http://ceciiil.wordpress.com/2008/08/03/bitter-soa/

    Do yourself a treat and watch the video from Martin Fowler and Jim Webb.

    REST seems to me a much better option for interconnectivity as it looks much simpler to deal with. But then we’re back to Andy issue regarding how open we want the access to content be. But that sounds quite a good idea, Oleg.

  • Ceciiil, I know, it looks like SOA is “yet another OOO or EJB or whatever else magics”… I agree, REST looks better, but only until you have a meaningful content. The biggest problem is that SOA, REST (and everything else coming) will die if your data (content) misbehaves in the same way as your spaghetti-C code before… My main point was, since PLM is all about data, we need to make it “service-oriented-delivered”. Nobody is doing it today as far as I know. Best, Oleg

  • Fully agree with your point Oleg about offering a flexible architecture able to provide easy-to-use content access services.

    However, PLM solution already are extremely complex (from both technical and functional perspective) and one should think twice before adding another layer of architecture complexity such as SOA.

    My point is about architecture complexity not about spaghetti code. Difference being that complex architecture can be implemented properly according to standards and still be slow, difficult to maintain and operate, deploy, customize etc …

    On the other hand, I believe that REST is a more natural, more straightforward and simpler (no WSDL etc …) approach to reach that goal of openness, interoperability and ease of resource access in PLM.

    Thnaks for the great post again. c.