Driving Toward an Open CAD Framework

Through the 1990s, technology had a major impact on individual disciplines and application tools creating higher and higher walls around the disparate functional and organizational silos found in product development companies. The technology became cheap enough to touch nearly every function and every task in each one of those silos across all organizations. Now, the most dramatic opportunity and most pressing need is to harness the technology to open and link the isolated silos through horizontal integration, in order to provide a consistent information architecture serving the complete product lifecycle across the full extended value chain.

PLM vendors such as Dassault/IBM, MatrixOne, PTC, SAP and UGS PLM Solutions recognize the opportunity to leverage web services to solve ongoing integration pain across value chains. Each has developed strategies to introduce this technology to enable its platforms to participate in, and perhaps orchestrate, cross-company business processes automated by these new capabilities.

Engineering and manufacturing processes, however, pose a challenge. The demand for increased applications performance in design engineering has established a particularly forbidding variation of “CAD Lock-In.” Emphasizing high performance and functional richness above nearly all other priorities, CAD application databases evolved highly specialized blends of hierarchical, relational, flat-file and object-modeling data structures, mixing and matching the approaches into unique results that follow none of the standard forms used commonly in other enterprise databases. They firmly marginalize CAD/CAM/CAE systems outside the enterprise mainstream and leave product development companies struggling to integrate their engineering data into a cohesive PLM framework.

Users currently articulate common pain points related to establishing a PLM environment independent of their industry vertical:

	Their information architecture is weak. Many users do not have a framework or architecture defined that maps across the enterprise, or supply chain. The demand for PLM solutions seen in the market today does not represent the potential available from full horizontal integration, backed by an enterprise data framework.
	IT and users cannot easily obtain needed data from PLM platforms. The lack of data standards and understandable formats compounds the challenge.
	Many users print information out of one system and retype it into another with “swivel chair” integration. Others employ the “Sneaker Net” with CDs to integrate CAD data. These manual pain points appear throughout the PLM process chain, but arose most notably with PLM, ERP, and legacy systems among nearly two dozen companies participating in our collaborative effort on integration last year.
	Implementing an end-to-end secure IT-based business process across applications and systems presents a formidable task today.
	Very few users understand the integration strategies of their own PLM and systems vendors.

Integration presents the overriding force pushing end user organizations toward open solutions – integration of processes, data, and application – or data access tools that support the flexibility needed for both the business and technical control they desire to manage and improve their product development process. Key to establishing a framework solution for this integration is attention to process. Data and software tools should be subservient to the role of supporting and improving process flows.

Despite its increasing importance, the term “open,” and its sister term “openness,” lack any form of substantive description that can be articulated and measured. While no one doubts the potential benefits that an open set of product development tools hold for users, the lack of a definition and measurable criteria for openness leaves both the developers of those tools and the users at a loss when they turn to focus on specific improvements. In the product design and manufacturing world, much of the discussion on openness has centered around the barriers in CAD products companies deploy to author product designs. All too often, CAD lock-in with proprietary databases isolates design and engineering. The common approach has been to narrow the focus even further, concentrating solely on the importing and exporting of native CAD database information used in those tools. While important, that focus invariably bogs down in long, drawn-out arguments attempting to establish agreement on a data format (both in content and semantics) for the exchange. In reality today, there remains a lack of consensus and an industry-wide risk that consensus will never be achieved.

User requirements for an open CAD framework fall into three high level categories: process, data, and tool dependencies. These needs extend well beyond the near exclusive focus on geometry creation by CAD vendors historically, through the 1990’s.

Process dependencies for an open CAD architecture center on the ability of CAD to communicate with both upstream business and technical applications that define product constraints and, then with downstream manufacturing and product support applications. For example, an up-front requirements analysis or systems engineering tool needs to inform CAD of engineering design constraints. On the business side, supply chain agreements negotiated by the company may be useful to guide the engineer’s choice for the selection of standard components within the product design. Downstream, manufacturing planning tools are used to stage and execute the production of the product models being developed in engineering. Often targeted to reduce time to market, these manufacturing planning applications, such as tooling, development of manufacturing process sheets, and assembly line layout evolve concurrently with the engineering design effort. They require controlled, timely access to design changes that occur in engineering. Moreover, the complexity and volume of data rises dramatically by two to three orders of magnitude in addressing downstream process requirements in manufacturing. Abstraction mechanisms become even more critical.

Independent of the actual data being accessed, dependencies for process integration center on access control and the ability to smoothly synchronize interaction between external applications and CAD. In a full open PLM environment integrated across the enterprise and supply chain, this synchronization forces requirements that even extends to the underlying technology that supports data communication between applications in the guise of web services and intra-network support.

CAD does not live in isolation. Open CAD data dependencies extend well beyond the issues with the integration of other engineering applications, such as analysis and simulation. Open CAD must communicate with enterprise-wide process control applications including workflow, change management, ERP, and supply chain management. Flexible solutions must address a broad range of technical concerns including access mechanisms, data granularity, and data transformation. Data access mechanisms import and export data from an application. At issue, the alternative demands for open CAD range from the data input/output for non-programming end users through the full flexibility of a robust programming interface for application developers. Clearly, the range presents very different ease-of-use demands.

Two extremes are external posting and a full application programming interface (API). External posting represents the ability to identify a specific data item (parameter, attribute, coordinate, etc.) that is then stored in a simple data structure external to the application. For example, values may be posted to a sequential data table or a spreadsheet. Also, the reverse access condition must be met – specifying a list of data values in a simple table or spreadsheet that may input into a CAD application. At the opposite extreme, an API access mechanism must support both read and write modes for the full granularity of CAD data content.

Data granularity refers to the level of application data that is accessible and how it may be collectively grouped for access. By its very nature, this question depends in the extreme on the specific application, as each creates and manages very different data. Centering on a CAD authoring as an example, data granularity includes at least the following:

	Product structure (including name, related Meta data such as bounding box and physical properties, parent/child relationships, and 3D transformation of instance)
	Geometry (includes both 3D exact and tessellated data, and 2D drawing and sketching data)
	Parameters and constraints (contextual data)
	Attributes (such as material, surface finish, color, etc.)
	Form features
	Annotations (including 2D drawing dimensions and notes)

The access semantics of external data must also be addressed.

The broad challenge facing the industry to establish an open CAD framework will not be solved until technical approaches emerge that provide the flexibility to support competing PLM solution vendors that will continue to pursue variations in their approach as they drive functional innovation and differentiation. Lessons learned with engineering solutions over the past decade must be adopted by the general PLM community at large. Abstraction mechanisms have proven to be critical in design, and now process abstraction should be explored. A single standard attempting to be universal, such as STEP, will take incredible time to gain momentum, and dramatically lags leading edge innovation. In addition, further research is needed in data architecture — the organizational structure of data across the enterprise, and the integration of CAD data into that architecture. Issues of a central data repository, versus distributed data, versus federated data architectures must also be resolved.

The Design Creation and Validation (DCV) service of Collaborative Product Development Associates, LLC (CPDA) is currently researching the requirements for CAD openness from an engineering perspective. A DCV multi-client Research Study and a collaborative User Council on CAD Openness have been launched for those interested in supporting the research and in developing flexible and agile “go-forward” technical strategies across product development environments. For more information on the CAD Openness Requirements study, and/or to request participation in the research, please contact a CPDA representative at sales@cpd-associates.com.

The results of the research effort will be presented at the CPDA annual conference, Product Lifecycle Management Road Map™ 2004, Sept.22nd and 23rd in Dearborn, MI. Join top analysts from Collaborative Product Development Associates, LLC along with key industry players as they share their experiences in making technology work by linking it across the full product lifecycle.