TechArchive

This article was originally written in the period 1995-2000

A philosophical piece from Bentley Systems, looking at how the structure of CAD companies has changed and continues to change, to reflect the products available.

In the early 70s, both CAD software and graphics hardware were typically developed by the same vendor and sold as a “turnkey system” by the vendor’s direct sales force. The architecture of such systems can be depicted as a series of concentric circles, representing hardware, CAD software and applications.

The vendor’s hardware in the centre included a general purpose mainframe or minicomputer. This system was complemented with additional graphics hardware for visualisation and graphics output, as well as user-interaction devices, such as a mouse or digitising tablet.

Around this hardware core, vendors often developed a proprietary CAD development environment that was used internally to develop vertical applications for the different domains.

Although this model was born out of necessity, because open systems were not available then, it did have several inherent advantages. First, since both the hardware and software were developed by a single vendor, this model did allow tight integration of applications to match the hardware and software architecture. Also customers were offered a single source for both hardware and software procurement and support.

But an inherent limitation of the “turnkey” model was that users were restricted to the applications provided by a particular vendor. These applications didn’t necessarily offer the best available for every application sector. Also, when graphics workstations became mainstream technology and open operating systems such as UNIX became more popular, uses wanted to run CAD on their preferred platform, rather than being restricted a proprietary one. Open systems were further enhanced by other software products, such as databases and logistics systems.

The migration from the closed and proprietary “turnkey” model towards a more open model occurred in the mid 80s. This trend was caused by two distinct developments:

• low cost graphics hardware platforms;
• industry direction to move from a direct selling sales model towards indirect distribution.

Lower cost, general purpose graphics systems started with UNIX vendors, but by the late 80s, both PC and Macintosh hardware had enhanced their systems with the graphics capabilities required to run CAD applications.

The model that replaces the “turnkey” systems model if often referred to as the “Virtual Company”. Instead of CAD being linked to a single proprietary system, it became available for all major hardware platforms. Also, by opening the CAD layer via Applications programming Interfaces (APIs), third-party developers are able to use the core CAD as the basis of their applications.

In the “Virtual Company”, the CAD vendor’s main responsibility is offer basic functionality in CAD and to work with hardware vendors to offer a portable environment. Third parties with expertise in vertical niche areas are encouraged to port to the CAD vendor’s platform and make applications available.

So the “Virtual Company” is based on synergy between the CAD vendor and the application developers. The application developers can concentrate on the area where they add value, without bearing the burden of developing and maintaining core graphics technology. The CAD vendor benefits by being able to sell the underlying CAD technology for every application sale, without the need to invest in development, support and sales to serve customers in vertical markets.

The “Virtual Company” is not only attractive from the points of views of vendors and applications developers, but it also offers some advantages for the end user. Instead of being limited to the applications provided by a single vendor, they can choose the application that best suits them. – for example, one that complies with local standards, or which is compatible with international requirements of the multinational company. Since there will be many application developers, they have an incentive to improve the products.

This model, which is the one commonly in use today, seems to work well. But there are some inherent limitations which affect technology, as well as marketing and distribution.

For example, the “Virtual Company” model concentrates on quantity of applications versus quality, ending up with a catalogue of several thousands of applications that an end user or reseller has to evaluate or choose from. Also, when a CAD vendor puts heavy emphasis on support, it is desirable to have the same standards for layered products, which is impossible to manage when dealing with thousands of independent developers. And from a developer’s point of view, a major limitation of the “Virtual Company” model is that it has proven to be very difficult for a smaller developer to sell applications beyond their own region.

But the major problem perceived by Bentley with the “Virtual Company” model is application interoperability. The use of CAD can only bring substantial productivity improvements if various applications can exchange data on a high level. When every developer bases applications only on core CAD technology, they can only share information on the low level denominator – geometry.

A geometry interface does not suffice when a piping application should be able to request a “hole” in a wall designed by an architectural application, or when both an AM/FM application and a civil planning application need to understand that something is a “street”, rather than a “line”. Interoperability is essential to integrate applications – and not only within one application sector, but across different domains.

The OpenSpace concept aims to solve these problems by having a core formed by a ring of horizontal products that run on various hardware platforms and provide an open architecture for customisation and development. These applications are all proprietary to Bentley and provide horizontal functionality for application domains at a higher level than with the “Virtual Company” model. Examples of these within the Microstation portfolio include TeamMate for document management, Descartes for raster editing and Masterpiece for photorealistic visualisation and animation. The products in this ring also include MicroStation 95, PowerDraft and Review.

On top of the horizontal tools, a series of sector-specific “integration platforms” provide basic functionality for a particular market sector and these can be used as the basis for third party developers to construct vertical applications with a high level functionality for a specific market sector. These platforms are referred to as a “Space” – for example, the mechanical engineering region is called “Mechanical Space” and so on.

The integration platform concept enforces certain standards that will enable third-party developers to exchange data on a high level. For example, by building applications using “Plant Space”, both a 2D schematics package and a 3D design application can consistently refer to the same object, “Pump”. Developers’ costs should be reduced, as the need to invent core technology from scratch is avoided by latching to the relevant integration platform. The approach also enables the integration platform to be optimised for core graphics and database environment.

The integration platforms are completely owned by Bentley, though in areas where the required expertise or experience is not available, they may be constructed using strategic alliances with a leading vendor in the relevant industry. An example is Jacobus technology, who will provide JSpace, an environment for Plant Design applications.

• Bentley Systems
• Tel: 01344 412233
• Fax: 01344 412386
• Contact: Sue Gale