This Website is not fully compatible with Internet Explorer.
For a more complete and secure browsing experience please consider using Microsoft Edge, Firefox, or Chrome

Analysis Origins - MSC and NASTRAN

Analysis Origins - MSC and NASTRAN

from the analysis origins series

by Dennis Nagy Principal, BeyondCAE

Analysis Origins - MSC and NASTRAN

Birth, Growth, Success, Stagnation, and Re-Birth

MSC’s 55-year history has been fascinating, exciting, and disappointing, all at the same time. This article attempts to highlight all those facets from the personal viewpoint of someone who was a part of it for 12 years (1985-97) and has followed MSC’s course closely since then. Like the famous cartoon of 5 blind men trying to describe an elephant from where they are touching it, this overview is a mixture of facts and personal opinions and I encourage any reader to contact me at dennis.nagy@beyondcae.com with comments, more detailed back-stories, or different viewpoints.

Contrary to what many people might assume today, MSC was not founded to develop a general-purpose finite element software system (NASTRAN), much less to be a global pioneer and, for considerable time, the market leader in the business of Computer-Aided Engineering (CAE). As the main co-founder, the late Richard H. “Dick” MacNeal once said in my presence (my recollection): “I started MSC because I didn’t like working for someone else in a large company.” Of course, the origins and motivations were more complex than that, but it’s a good summary of how MSC started.

Figure 1: Richard H. MacNeal 1924-2018

 

Figure 1: Richard H. MacNeal 1924-2018

MSC (called The MacNeal-Schwendler Corporation at that time) was incorporated on February 1, 1963, by Dick MacNeal and Robert “Bob” Schwendler. Caleb “Mack” McCormick was with them from the early years but didn’t get his name in the company name. Dick MacNeal was the theoretical/scientific driver for starting MSC as a company to develop analog computer technology for, among other things, helicopter rotor dynamics, and to do consulting related to that. These three pioneers and others, however, saw the coming demise of analog computing and the digital computer revolution on the near horizon. Through various aerospace industry and NASA contacts, they were able to team with Martin Baltimore and Computer Sciences Corporation (CSC) to submit a winning bid in late 1965-early 1966 to NASA to develop a very comprehensive (by mid-1960s standards) finite element analysis (FEA) software system. Hence NASTRAN (NASA Structural Analysis) was born.

The intervening years from the mid-1960s to early 1980s were filled with fascinating developments and details. Unfortunately, that was before I was part of MSC but, fortunately, Dick MacNeal chose to write an excellent book entitled “The MacNeal-Schwendler Corporation: The First Twenty Years” upon his retirement from fulltime company management and on the occasion of MSC’s going public (on the American Stock Exchange) on May 5, 1983. Dick’s excellent writing style, coupled with his unequalled wealth of detailed personal recollections, produced a unique book which I cannot even begin to adequately summarize here. There is, to my knowledge, no other summary of MSC’s history shorter than that book but more detailed than this present article, so any reader interested in those fascinating years must read the book, which I strongly encourage you to try to do (if you can find a copy in print—I have one signed by Dick MacNeal on April 13, 1988 and there are some others floating around among former MSC employees). There is also one other brief historical summary, vintage late 1990s and viewing MSC’s future prospects from that vantagepoint [1]. I will jump quickly to the period 1985- 2018 for the remainder of this article, after briefly covering the history of MSC’s legal dispute with NASA, how MSC’s pricing model sustained financial growth up to and well beyond 1985 and how MSC entered the Supercomputing era with MSC/NASTRAN on a Cray.

Richard MacNeal (right) and Robert Schwendler

 

Figure 2: Dr. Richard MacNeal (right) and Robert Schwendler in the early 1960s working on an analog computer.

The birth of a proprietary version, called MSC/NASTRAN, occurred in 1971 because NASA had not provided for any ongoing user support, error correction, and further enhancements for the public-domain (NASA) version of NASTRAN. Since, on day one, the actual MSC/NASTRAN software was identical to what anyone could obtain from the NASA COSMIC distribution center at the University of Georgia for a modest fee of $1,750 (mostly tape-copying and shipping fee), MSC could not charge any significant price for a perpetual (paid-up) license of the MSC/NASTRAN software itself. Hence the idea of charging a monthly fee (which became known as a lease fee) for “hot line” telephone support, ongoing bug-fixing and enhancements, better documentation, and user training courses was used to start generating revenue for MSC. It proved successful enough that MSC grew well throughout the 1970s without any need for outside investment. It is worth noting that MSC held onto primarily a lease-based pricing model for decades while other software companies needed to resort to frontloaded paid-up licensing to generate enough early cashflow for growth. Much more recently, major successful software companies have introduced “subscription pricing” to replace perpetual licenses, as if this were a revolutionary new idea.

There was one legal “hiccup” with NASA during the period 1980-1982 that is worth brief mention here, although covered by a whole chapter (Chapter Ten: Dispute With NASA) in Dick McNeal’s book. Someone (later known to MSC and Dick MacNeal but not mentioned directly by name in the book) had formally complained to NASA about MSC’s “misuse of NASA data” without prior NASA approval, and of course NASA had gotten its legal department involved. If you love to read about gory legal details, read Chapter Ten. Here I will just summarize a long list of painful (for MSC) iterations and corresponding legal fees that ensued (even noted Pulitzer Prize-winning U.S. political columnist of that time, Jack Anderson, had written about the MSC-NASA issue in February 1981). The need to reach a negotiated settlement was quickly clear to both sides.

Many of the later iterations were about what MSC should pay to NASA/the U.S. Government to settle the dispute. Initial (posturing) amounts ranged from a few thousand U.S. dollars up to US$900,000. After a lot of careful and admirably detailed calculations on both sides, MSC and NASA arrived, on October 22, 1982, at a deal of US$125,000 and the right for MSC to use the entire software it had developed, and the name MSC/NASTRAN, from that point forward. (Joe Gloudeman later referred to this settlement as a “Quit Claim Deed”.) Even at that time, and much more obvious now in retrospect, that amount was seen by many as “peanuts” compared to what MSC/NASTRAN could (and did) produce in revenues and profits over the ensuing 3+ decades. A related issue to this apparently settled dispute did, however, appear in in 2001 and is summarized further below.

Robert Schwendler

Figure 3: Robert Schwendler, with MSC’s first superminicomputer (VAX 11/780), late 1978, shortly before his untimely passing

In addition to entering the super-minicomputer market with an MSC/NASTRAN version on a Digital Equipment Corporation (DEC) 11/780 in late 1978, MSC’s most significant “porting” of MSC/NASTRAN was to a new class of vector-architecture supercomputers via the creation of a Cray version in 1979-81. At that time of commercial supercomputing infancy, MSC and Cray estimated that, at most, approximately 25 Cray computers would ever be sold globally by Cray for running MSC/NASTRAN. That original estimate was already exceeded by at least a factor of 10 during the next 10-15 years! The cost of Cray computers meant that only the largest MSC customers in the Aerospace and Automotive industries, as well as some NASA data centers, could afford them. MSC’s software pricing was, at that time, linked to measured usage of compute cycles (and the hardware cost of such cycles) in order for MSC to get a piece of the economic benefit from large mainframe- and supercomputer-based “time-sharing” data centers. Ironically, the Space Shuttle Challenger disaster in 1985 eventually triggered the slow unraveling of MSC during the 1990s. After the Challenger explosion, NASA’s subsequent intense MSC/NASTRAN usage (on Craybased data centers) to help understand what caused the failure led to a significant revenue windfall for MSC due to such usage-based pricing over the next 4 years.

The Beginnings of Stagnation and Two Decades of Turmoil

That windfall, in turn, gave MSC a somewhat false sense of financial confidence (and the stock market a false sense of MSC’s growth potential) which partially led MSC to start acquiring small complementary CAE software companies with offerings in subject areas where MSC/NASTRAN was difficult to enhance. These acquisitions were not easy to absorb culturally/organizationally and hence were a drag on MSC’s bottom line. When the windfall stopped blowing in 1989, MSC’s bottom line started to suffer measureably, leading (with some other extenuating circumstances) to MSC’s Board removing Dr. Joseph “Joe” Gloudeman as CEO in September 1991 (he had been running MSC since 1984 when Dick MacNeal had retired) and bringing back Dick MacNeal to run the company. At age 67, Dick MacNeal viewed his return as a temporary move and proceeded (with the Board) to look actively for a new CEO.

In the mid-1980s MSC recognized that the 1960s-vintage foundation architecture (the Executive System [ES], for handling all module and data control) of MSC/NASTRAN was no longer an adequate base for longer-term enhancements and extensions. A complete re-write of the ES was performed during 1985-87, took longer and cost more than originally estimated, but was successful enough to make MSC/NASTRAN a more viable development platform for decades to come.

As affordable graphics hardware and workstations/servers emerged, market interest broadened to require more powerful pre/ post-processing tools (interactive preparation of finite element models, meshing, and visual portrayal of results). MSC struggled with its own in-house developments in this area (MSC/Grasp and then MSC/XL) from 1983-1991 before deciding to acquire Massachusetts-based Aries (privately held developer of FEA graphics and related solid modelling software) in 1993 for ~US$15M and then (because it suddenly became available) publicly-traded PDA Engineering (PDA/Patran) for ~US$60M in mid-1994. PDA/Patran and SDRC’s I-DEAS were at that time the coleaders in FEA pre/post-processing software and both were used heavily by MSC/NASTRAN. PDA/Patran, originally written in the late 1970s, was completely rewritten in 1992-94, ironically a reason for PDA’s earnings and stock-price collapse in early 1994, leading to their “availability” to be rescued by the Dick MacNeal-driven deep pockets purchase.

The major acquisition of PDA in 1994 coincided with a CEO “musical chairs” that MSC experienced going forward, which in turn played an important role in MSC’s stagnation and slow unraveling. After Dick McNeal’s 2- year return as CEO (1991-93), MSC chose Larry McArthur (CEO of acquired Aries) to succeed him in 1993, removed MacArthur in 1994 (at the urging of still Board Chairman MacNeal) after the PDA acquisition (which in hindsight made the Aries acquisition appear superfluous, to this author at least, except for the money spent on acquiring it) and then chose Tom Curry (PDA’s President) to run MSC in 1996. Recognizing the precariousness of his new role, Tom, with the aid of the former PDA and now MSC Board member Frank Perna, convinced the Board to remove Dick MacNeal from the Chairmanship and the Board in 1997. This didn’t lead to the stability Tom had hoped for though and at the end of 1998 Frank Perna replaced Tom Curry as CEO.

A Period of Difficult Acquisitions

The Frank Perna era (1999-2004) was characterized by several ambitious acquisitions and (generally unsuccessful) attempts to broaden MSC beyond a pure FEA-based MCAE leader MSC acquired the two small (but annoying due to their low pricing) other NASTRAN vendors (who had built their NASTRAN products from the same public-domain source code MSC had developed for NASA) in 1999. These acquisitions were to come back and bite MSC financially a few years later (see below).

Privately-held MARC, a co-leader (along with HKS’s Abaqus) in the nonlinear FEA market niche, was acquired in 1999. MSC had struggled throughout the 1990s to implement nonlinear capability in MSC/NASTRAN, which was basically still a linear-FEA architected software system. Nonlinear FEA was becoming more important to MSC’s automotive and aerospace customers because the rapid decrease in price-performance ratios for relevant processing hardware made useful nonlinear simulations much more practical in industry. At that point, in addition to MARC and ABAQUS, Ansys also had better nonlinear simulation capability than MSC/NASTRAN. Although a wise acquisition in principle, the other management/strategy problems MSC had from that point onward produced the result that MSC.MARC (somewhere in there MSC switched from / to . in its product naming convention) was only really integrated with MSC.NASTRAN 13 years later.

Attempts to broaden MSC led to the acquisition of AES (a primarily U.S. active DS/CATIA and bundled workstation hardware re-seller/implementation consulting house) for ~$100M in 2001. Many factors (too speculative to discuss here in writing) led to MSC’s eventual sell-off of the remains of AES a few years later for a small fraction of what they paid. In the same timeframe, MSC also attempted to configure and sell compute servers based on an MSC-created version of Linux, another venture with much greater start-up costs than eventual revenue, which was abandoned within 2 years.

Perhaps the best (from a complementary technology and strong, overlapping market presence viewpoint) acquisition MSC made during that period was the purchase of publicly-traded MDI (multibody dynamics [MBD] simulation called ADAMS) in 2002 for slightly more than US$120M. MDI, at ~US$60M in revenue, was by far the leader in the MBD segment, but a rumored acquisition battle with Ansys may have (again after the fact speculation) caused MSC to pay too much for MDI. The combined organization after that acquisition was ~1700 people. In hindsight, MSC.ADAMS and its dominant MBD market segment share has played a key role in retaining the loyalty of some of the same automotive customers battered by the gyrations of MSC in the “NASTRAN” market. MSC.NASTRAN was used for “bread and butter” linear FEA simulation of large automotive components and sub-systems and many customers realized that they could reliably do the same simulations with other “lesser and cheaper” NASTRANs or even with non-NASTRAN FEA (mainly Ansys, Abaqus and Altair’s Radioss more recently).

MSC was clearly stagnating and slowly unraveling in the early 2000s. To make matters worse, in 2001 the Federal Trade Commission (FTC) hit MSC with a claim (called an “administrative challenge”) of monopolistic practices in the “NASTRAN market” for acquiring the two other, much smaller NASTRAN vendors (UAI and CSAR, both Los Angeles-based) back in 1999. The challenge was rumored to have come from a complaint to the FTC by MSC’s largest aerospace customer at the time (over US$4 million in annual lease/usage payments to MSC) related to their fear of monopolistic future pricing. The FTC’s challenge was persistent despite many MCAE experts, including this author, providing testimony to the FTC that there was an FEA market, but no longer any “NASTRAN” market at that time due to the continued growth of both Ansys and HKS/Abaqus. MSC ended up spending ~$7M in legal services and still lost the challenge in 2002 and were forced to “divest of their monopoly” by offering for sale (at a “reasonable” price) a clone of MSC.NASTRAN source code and full documentation, plus allowing the chosen acquirer to hire a sufficient number of key developers to be competitive with MSC going forward. Unigraphics (UG, a major MCAD vendor with its NX-CAD environment) ended up buying the clone, productizing it as NX/NASTRAN and marketing it at a lower price than MSC.NASTRAN.

MSC’s Board removed Frank Perna as CEO in 2004 and brought in William “Bill” Weyand, the former SDRC CEO who had successfully sold SDRC to Electronic Data Systems (EDS). Outside rumor was that Bill Weyand would have four years to improve MSC’s stagnant financial performance and find a buyer to take MSC private. MSC continued to be revenue-flat and Bill Weyand left at the end of a four-year term. The MSC Board inserted a temporary placeholder CEO while directly looking for a buyer, which they found (STG: Symphony Technology Group, a private equity firm) in mid-2009 at a purchase price of US$360 M.

The Beginnings of Rebirth

STG subsequently hired CAD Industry veteran Dominic Gallello as CEO, who returned MSC to break-even and embarked upon an ambitious, aggressive development plan starting in 2011 to replace MSC.PATRAN (now 18 years old in its underlying architecture and surviving during a period of dramatic software methods improvement within the industry). MSC.APEX was the result of this ambitious project and seems to be making reasonable headway today.

As is the case with many successful private-equity companies, they have an approximately 5-year ROI horizon. STG attempted to sell MSC in mid-2014 but was unable to find a willing buyer at the price they wanted to achieve. STG attempted again in 2017 and successfully sold MSC to Swedish-based Hexagon AB, a leading global provider of industrial information technology software and services, for US$834 M. Although MSC had made a number of technology-attractive investments and acquisitions in the period 2010-2017, they had insufficient impact on MSC’s top line so that MSC’s revenue had languished in the low US$200 M range for many years prior to this successful acquisition. It is of course too early to tell, but this acquisition has the potential to get MSC back onto a more aggressive growth path with the synergy of Hexagon, a major industrial player instead of a private equity company, as its new parent.

References

1. The MacNeal-Schwendler Corporation History: Funding Universe 1999 nafe.ms/2OmdSm2
2. The MacNeal Schwendler Corporation: The First Twenty Years, Richard H. Macneal, 1998 nafe.ms/2OkOPzY

Posts By Month

Tags

Analysis Origins

All Tags


blog

Author

Dennis NagyDr. Dennis Nagy is a 47-year veteran of international experience and accomplishments in all aspects of the global engineering simulation (CAE) software industry, including in-depth experience in building high performance teams in multi-cultural / multi-location environments. He currently provides global strategy, business development, and industry/market analysis consulting to multiple software companies and private equity/VC investment firms, as Principal of BeyondCAE. He is currently particularly focusing on the Democratization of Engineering Simulation and the key role Simulation Applications (SimApps) and Digital Twins will play in the dramatic expansion of business-beneficial engineering simulation use.

Prior to his current consulting practice, he held various executive positions, including VP of Marketing and Business Development at CD-adapco (now part of Siemens PLM), CEO of Engineous (now part of Dassault Systemes), Senior VP of Worldwide Sales and a corporate officer at publicly-held MSC.Software (now owned by Hexagon AB), VP of Marketing and Asia-Pacific at Fluent (now part of ANSYS), and VP of International Business Development at Blue Ridge Numerics (now part of Autodesk). Earlier in his career he ran the CAE organization at GE Medical Systems and opened the German office of SDRC (now part of Siemens PLM). He was also an Assistant Professor of Engineering at Princeton University and a Research Engineer at the University of Stuttgart’s prestigious Institute for Statics and Dynamics of Aerospace Structures (ISD).

Dr. Nagy is an engineering graduate of MIT (BS, MS) and the University of California, Berkeley (PhD). dennis.nagy@beyondcae.com

 

View all blog posts