Architecture Matters

We live in a time of unprecedented disruptive innovation—the destruction of those unable to compete and the emergence of smart, connected products and services. At the heart of this continuously accelerating race to innovation is the need to support the development of increasingly complex multidisciplinary products with electronics, embedded software, internet access and configurations throughout the product lifecycle. While the PLM vision continues to expand to meet these needs, the state of PLM implementations as reported by CIMdata has not kept pace.

And yet, as we have reported on extensively in recent years, the scope of PLM implementations has not changed much over this time. Hard to implement, expensive to upgrade, and too focused on managing mechanical CAD data, many of these solutions have failed to meet the expectations and needs of their users, especially in large organizations. The result is widespread use of complex spreadsheets, email, and bespoke databases to fill in the gaps among legacy systems. (CIMdata, July, 2017)

Many established industrial companies base their product development on PLM systems that rely on monolithic architecture and a series of acquired products built on top of different architectural frameworks. They may use terms like “backbone,” and “platform” but coupling together different architectures does not constitute an actual platform.

There is an excellent article that appeared in the Harvard Business Review by Mark Bonchek and Sangeet Paul Choudary, entitled, Three Elements of a Sucessful Platform Strategy that provides a good background of what a platform strategy is. CIMdata uses this work in their position paper on Product Innovation Platforms, Their Role in the Enterprise, and Thier Long-Term Viability, which describes in detail all of the elements of a Product Innovation Platform.

A platform is something that is foundational, that you can build upon that is “plug and play.” As it relates to Product Innovation Platforms, I would like to focus on the one aspect that I think matters the most—Architecture.

I contend that more functionality in a given area is useless, if you do not have an open, flexible and sustainable platform that allows you to take advantage of it. If you cannot “plug and play” and seamlessly stream product data across your Digital Thread, then all you have done is added another disconnected island of functionality. These chunks of monolithic architectures “end up looking like pools of mud” (Simon Brown).

Architecture matters. Unfortunately, within the current state of PLM systems, the architecture is quite poor. First, they are not based on an open consistent end-to-end resilient architecture, and second, they perform customizations made to the applications layers are that are hard coded compiled software to rigid architectures, which are often “proprietary monolithic” architectures. This results in lengthy implementations—some lasting years with time consuming upgrades or, worse, no upgrade at all. They fall years behind until a migration is the only option. By ignoring the importance of the underlying architecture, companies cannot pivot in hours to days but months to years.

Resilient Architecture

A resilient architecture gives you the flexibility to change quickly and endure over time, which is a fundamental element in all winning organizations. The only thing you can predict with certainty is that there will be disruption, both in the marketplace and the technologies required to compete. Therefore, a resilient PLM Platform that can accommodate future changes with minimal disruption is the key. If your PLM system is a series of architecturally distinct products, then you DO NOT have a resilient architecture, nor a Product Innovation Platform. This leaves you vulnerable to expensive customizations, upgrades, suboptimal processes, and at a much higher total cost of ownership. Worse, it leaves you with an inability to respond that could be catastrophic in terms of getting innovative products to market, improving operational efficiency, and increasing your customer engagement.


CIMdata defines “Openness” as one of the keys to a successful Product Innovation Platform, which they define as: provides unencumbered access to product-managed data, workflows, and services; openness ultimately means both plug and play and transparency without using proprietary, “monolithic” architectures.

 The key to flexibility is having an open architecture, allowing a company to control its data, its workflows, etc. It has been quite common for legacy PLM solutions to use compiled logic and proprietary data models, then customers have to customize on top of that, only to find it changed in a new release or objects or triggers previously used are no longer supported. These nightmarish scenarios are all too common with many existing PLM legacy tools.

Your data belongs to you, which seems obvious, but if you can’t easily extract it or are forced into the vendor’s schema change, which may conflict with previous customizations, then your data is being held hostage. Aras enables every company to access, use, customize, integrate, export, and move the data they generate, including the processes, data models, business rules, and system definition they create. To enable true interoperability, portability, and extensibility, your PLM should be based on an open architecture featuring an open data model and open interfaces. Control your own destiny and do not be locked into proprietary schemas. A modern open architecture provides you with the flexibility to be prepared for tomorrow.


Many PLM systems struggle to scale due to limitations in their architecture, including reliance on a single complex database schema. Make sure the architecture scale out and scale up to accommodate you and your value chain.


CIMdata cites “Upgradability” as another key to a Product Information Platform. Companies need to know they are customizing on top of an architecture that is inherently upgradeable.

Aras Architecture

The Aras architecture is model-based modularized architecture where the applications are driven by an independent underlying service. This delivers an agile innovation layer, providing companies with an agile Product Innovation Platform that connects disparate systems and eliminates the hard coding and recoding of object behaviors found in legacy PLM architectures. The Aras architecture is comprised of four layers: Client, Application, Services, and Repository. The Client layer is where users access the platform via a web-based interface, mobile app, MS Office, or other integrated tools. The Application layer is powered by the modeling engine, which manages the PLM applications by subscribing to the PLM behaviors defined in the Services layer. Aras uses meta-data templates to describe everything: application functionality, business logic, data schema, screens, workflows, etc. The meta-data templates define the objects in the system and those objects then subscribe to the services they need in the Service layer. This modular architecture is the basis for creating flexible, scalable, and upgradable applications. Aras build its PLM applications with the powerful and flexible modeling engine, just as companies use it to write their own customizations.

The Services layer is where the core PLM objects, workflow, etc., reside, and the real work of PLM is done. The Repository layer is where the application models and vaulted files are stored, which can be packaged and moved at any time.

Instead of legacy PLM's approach of defining an object model and hard coding it into the services, the Aras Product Innovation Platform separates the two, making it resilient. The result is a run-time web application consisting of loosely coupled/federated web services optimized for performance and scalability to form a resilient modular architecture.  This allows Aras to upgrade our customers’ customizations.

The architecture utilizes a dynamic schema with loosely coupled/federated web services that enable scale-out on the file system and Web server and scale-up on the database by moving non-transactional data to scale-out servers. This allows Aras to handle large levels of throughput by using frequent, short database transactions. Aras has been independently tested up to 250,000 concurrent users with excellent response times. Aras is not constrained by its architecture, so, by design, it puts the burden of work on the hardware so it can scale-up and scale-out.

The challenge that companies face is how to handle the constantly growing and changing complexity of the products we develop. The opportunity is to realize that an open, flexible, scalable, and upgradable platform built on resilient architecture makes complicated things sustainable. A strong open architecture enables flexibility, and flexibility is the key to a Product Innovation Platform that can quickly connect applications and data across the product lifecycle as well as respond by quickly changing applications, processes, and customizations—all of which are easily upgradable. The world does not stand still. Tomorrow’s requirements need to be met quickly in order to be a leader and disruptor in the marketplace.

Shape your future before it shapes you.