Siemens in the business of supplying hearts and parts to every other CAD vendor.
Cambridge, U.K. is the home of Siemens PLM’s components group, responsible for Parasolid and D-Cubed, makers of components in use by many popular CAD products. Parasolid is a widely used geometry kernel. D-cubed’s
2D DCM and 3D DCM are constraint managers. Siemens licenses these components to practically every CAD company—in addition to using them in Siemens’ own NX and Solid Edge products. This makes them the most widely used CAD components in the world.
Selling components to a competitor makes for bizarre relationships. Imagine a car company supplying engines to every other car company. It’s been going on for years and the CAD vendor community seems to be oddly comfortable with it. SOLIDWORKS and Onshape, arch-competitors to Solid Edge, both use the Parasolid kernel and have never shown any desire to develop their own. We hear from Siemens that the company stands to profit more from licensing its components than if it would by keeping Parasolid exclusively for its own purposes. That seems doubtful when you consider that a competitor would cease to exist with its Parasolid heart ripped out and Siemens products would take over the dominant position in a product design software market worth billions.
Engineering.com was in the neighborhood and sought to visit the Siemens facility in Cambridge. But we were told everyone we wanted to see was elsewhere and unavailable. Would we mind sending our questions via email. The result follows:
When did Parasolid first appear?
Neil Howarth, Business Development Manager of D-Cubed Components. (Image courtesy of Siemens.)
Parasolid v1.0 was released by Shape Data Ltd. in 1988 to supersede its ROMULUS product, which was the first commercial b-rep modeler launched in 1978. The principal differences were that ROMULUS was single-precision—supporting analytic-only geometry with a FORTRAN-style API—whereas Parasolid was double-precision—supporting analytic and NURBS geometry with a C API. Models from Parasolid v1.0, and even ROMULUS, can still be loaded and used in the latest version, Parasolid v31.0
What are Parasolid’s advantages over the competing products?
The fundamental advantage is the Parasolid culture which has always been focused on solving the most difficult customer geometry problems through innovation and meticulous software engineering. Following from this is recruitment of the best talent, building long-term customer relationships and investing in processes to optimize the breadth of functionality, performance and quality of our product. This is witnessed by numerous technology firsts we have brought to market, including tolerant modeling, two-way data compatibility, support on iOS devices and, currently, Convergent Modeling. Parasolid’s business proposition is also designed to encourage innovators, large and small, to experience our technology and customer support before entering into a win-win relationship.
What percent of the world’s solid models are being created with Parasolid?
It’s getting more and more difficult to estimate this accurately. However, we license Parasolid to over 130 independent software vendors and estimate (conservatively) that we have 4,000,000 end-users doing heavy-duty model creation and editing each day—that is, using Parasolid-based applications from companies such as SoldWorks, Onshape, Vero, Missler, Vectorworks, Bentley, Fujitsu and, of course, Siemens PLM (NX, Solid Edge, Tecnomatix, SimCenter). Perhaps two to three times that number of end-users are using Parasolid-based applications for less demanding modeling, view markup, interoperability and similar tasks.
The Parasolid XT file format is published as part of the ISO JT specification and supported by the JT Open Toolkit. This means that many millions of models containing precise geometry compliant with the Parasolid XT definition are created and shared globally throughout industry.
With ACIS fading, is Parasolid close to having a monopoly?
In terms of Parasolid licensees, we have a leading position in the market, but we don’t see any elements of a monopoly situation developing. We frequently compete with other kernel modelers in sales situations and are keenly aware of emerging technologies and market shifts that present new challenges, e.g. generative design modelers. What we do see are lots of opportunities to make Parasolid more powerful in order to assist Parasolid-based licensees to become more competitive in their markets.
Does Parasolid handle scans/triangulated surfaces/point clouds?
Before answering this specific question, a little background is helpful. Several years ago, we first saw facet data entering engineering workflows in a meaningful way, as high quality scanners became easily accessible and medical scanners became common place. At about the same time, our customers began expressing interest in Parasolid’s strategy for supporting facet data as they developed their own facet support strategies.
After extensive prototyping, we decided to fully integrate facet data into Parasolid’s data model and algorithms as a new surface type alongside analytics and NURBS. This is the most elegant technical solution, but also the most challenging to implement, as it involves enhancing every algorithm that handles surfaces. But, more importantly, it provides the best architecture for the long term to deliver seamless functionality, minimal integration costs, low performance overheads and the levels of quality Parasolid customers expect.
We call this new technology Convergent Modeling because it converges facet modeling technology seamlessly into a B-rep modeler for the first time. The end-state of Convergent Modeling is that all Parasolid operations will support models containing arbitrary combinations of classic B-rep geometry and facet B-rep geometry. In the current version of Parasolid, models can either be 100 percent classic B-rep or 100 percent facet B-rep and the vast majority of operations already support 100 percent facet B-rep models. Our customers can test this by converting any classic B-rep model into a facet B-rep model with a single function call. Architectural work is well advanced to enable operations on mixed models, containing both classic B-rep and facet B-rep geometry, and this functionality will be rolled out in sensible, usable chunks over coming releases.
It is important to stress that Parasolid is still focused on classic B-rep modeling and that resource has been added to implement Convergent Modeling so that the pace of classic B-rep enhancements is not impacted.
To address the original question, Parasolid works with imported facet (triangulated) data via Convergent Modeling technology. For example, STL files can be loaded, analyzed, repaired as necessary and then used as Parasolid surfaces on models.
Parasolid doesn’t currently support import of point cloud data because there are many solutions already available and many of our customers have their own code for this. As such, it is not a high priority as we drive to deliver the Convergent Modeling end-state.
How does Siemens justify licensing its technology to competing companies?
[This is written in terms of Parasolid for simplicity. Much of the information is similar but sometimes subtly different for our other components]
Siemens Product Lifecycle Management (SPLM) has always aimed to be a helpful, constructive, collaborative partner enabling end-users, large and small, to help them gain competitive advantage through mutually-beneficial, long-term relationships. Licensing our component software over decades has supported this philosophy in a number of ways:
- Data Openness – End-users are more comfortable knowing that their model data is stored in the published Parasolid XT file format rather than being trapped in a proprietary and/or encrypted format. The XT format is also part of the ISO JT Open standard specification and many large manufactures who are not SPLM end-users choose to archive their models in XT format within JT files e.g. Ford, Caterpillar, Volkwagen-Audi (see graphic above).
- Heterogeneous Environments – SPLM end-users are able to select best-of-breed applications from over 130 independent software vendors (ISVs) to mix and match in Parasolid XT-based workflows. If SPLM does not have a suitable offering or the best offering, the end-user should not be disadvantaged by limited choice.
- Better Components – Parasolid is improved by working with a huge diversity of industry experts and innovators from ISVs worldwide. Similarly, through licensing, Parasolid has involvement in a wider diversity of application markets. For example, SPLM is not present in the Architectural, Engineering and Construction (AEC) market, but Parasolid is the core modeler of Bentley, Vectorworks and Allplan products.
- Building Relationships – Over several decades now, SPLM has proved itself as a trusted supplier of leading technology with the interests of our licensees at heart. This is exemplified by the Level Playing Field policy that promises the Same Software, at the Same Time with the Same Support. This is to reassure licensees that, when they build their business on Parasolid, SPLM will not use its ownership as a competitive lever. Many of our licensees have been with us for over twenty years and several industry leaders are serial licensees of Parasolid for different companies and start-ups.
- Opportunities – having respectful relationships with a large proportion of the PLM supplier community makes things happen. It may be a simple M&A proposal, a reseller agreement, a joint venture or something industry changing like the JT Open ISO standard.
When did D-Cubed first appear?
Peter Kerwin, Product Manager, Parasolid. (Image courtesy of Siemens.)
D-Cubed Ltd, the original company, was founded in 1989. D-Cubed 2D DCM (Dimensional Constraint Manager), the original geometric constraint solver, was first released in 1990.
What are D-Cubed advantages over the competing products?
There are many factors including:
- The length of time for which the D-Cubed components have been developed. We were first to market by a long time.
- The total investment over time, and the current level of investment. D-Cubed 2D DCM has a dominant market share (see the final question), and the income this generates is reflected in a dominant investment position.
- If you have a dominant market share, and have been around for nearly 30 years, then you’re more likely to be around for the next 30 years. Competing components have a much shorter track record, and far fewer resources to support ongoing developments. So, long term security of supply is a big D-Cubed advantage.
- Dominant market share and investment levels lead to broader functionality, faster performance and higher reliability than competing systems.
- We’re located in Cambridge, U.K., a leading global center for technology and a magnet for fantastic software developers.
What Percent of the world’s solid models are being constrained with D-Cubed?
This is impossible to say. Measuring market share as the percentage of revenue from the licensing of geometric constraint solving components, or the number of end-user seats based on geometric components, then we estimate the D-Cubed 2D DCM market share to be over 95 percent. Then there are CAD systems that are based on in-house geometric constraint solvers, not on geometric constraint solving components, perhaps used in somewhat less than 50 percent of end-user seats. Very roughly you could estimate the percentage of the world’s solid models being constrained by D-Cubed 2D DCM as 50 percent.