Although I retired early from the space business in 2004, I feel compelled to answer the question posed by Loren Thompson (no relation): “Why is MDA Moving So Slowly on Space Sensors?” [Commentary, March 21, page 27]. In a badly misinformed op-ed, Thompson tries to make a case that the Space Tracking and Surveillance System (STSS) missile tracking technology should be produced and deployed immediately instead of the very wise decision by the U.S. Missile Defense Agency (MDA) to kill STSS and look for something that works.

The sad truth is that the technology deployed aboard STSS would not work effectively in many scenarios. The technology is better suited for the Smithsonian, where it was headed before being rescued in a fateful decision by Pete Aldridge, former undersecretary of defense for acquisition, technology and logistics, over eight years ago. MDA subsequently made the right decision to scrap the STSS space junk and to try to move on to a solution that will work.

As for MDA’s decision to sole-source the next job to Johns Hopkins, the “Space Earmark University” — well, for that new adventure there can be few words. Maybe Hopkins can get Big Aerospace to get it right after reinventing the wheel for the fourth time and adding even more layered costs.

From 1999 through 2003, my former company, Spectrum Astro Inc., competitively won and ran a $285 million U.S. Air Force research and development contract for STSS under its previous name, Space Based Infrared System (SBIRS) Low, going head-to-head against TRW (now owned by Northrop Grumman). The only way the performance assessment can be done on a system like SBIRS Low is through very complex global simulations and thousands of “Monte Carlo” runs, or repeated random sampling. When we won the contract, the Air Force handed us a copy of its low Earth orbit missile detection system simulation software, which had been developed by a Beltway bandit that had been given $10 million over several years to develop a SBIRS Low simulation. We bought the very latest and fastest Silicon Graphics computer hardware to run it on.

Within days of running realistic simulations of even limited attacks on the United States, both the Air Force software and the Silicon Graphics machine were brought to their knees. Trying to run even one Monte Carlo simulation took 45 minutes and repeatedly crashed both the software and the machine. We needed 10,000 Monte Carlo runs to get the statistics needed. We decompiled the Air Force machine code back to source and realized that we had been given spaghetti code with all kinds of erroneous and “plug number” assumptions. We hooked a log chain around the Silicon Graphics machine, yanked it out of there, turned it and the software into a world-class boat anchor, and started from scratch.

The job to develop the new comprehensive simulation fell to Allison Floyd, one of the brightest programmers I have ever met, and a team of equally bright young hacks around her. It was Nerd ’n’ Geek Heaven. We shoveled in the cash and whatever computer hardware and software they asked for. Facilities and information technology crews were assigned full time to build, move or modify anything they wanted. We piped in sunshine and they all ran 24/7. Within about 90 days, Floyd and her team had developed a brand-new, fully interactive global simulation of many SBIRS Low satellites complete with sensors flying across seven independent moving and rotating reference frames. We were up and running on over 1,000 parallel Intel Pentium microprocessors running at speeds up to 5 teraflops, one of the largest supercomputer arrays in the U.S. Southwest, all ordered over the Internet and delivered to us by UPS.

Hunkered over dozens of blue screens in darkened (and freezing) secure rooms behind at least three cipher locks, and surrounded by empty pizza boxes and cases of Diet Coke, within nine months the team could “launch” one-to-many full-fidelity simulated missiles carrying one-to-many full-fidelity warheads from anywhere on Earth to anywhere on Earth, and assess the ability of any array of “flying” space sensors in any constellation to detect, track, discriminate, analyze and report the attack to somebody on Earth that could do something about it. Single launches, “tube raids” (multiple missiles flying close together as in the shape of a soda straw), “fan attacks” (where several missiles diverge like the tines of a leaf rake near the end of flight), “depressed trajectories” (where missiles are flown at lower altitudes to make them harder to detect) and many others were flown over and over. Many of the scenes in that room could have been right out of the movie “War Games.”

In the midst of all that, professor Ted Postol of the Massachusetts Institute of Technology launched his own attack on the viability of ballistic missile defense. Working silently, we modeled and evaluated each of his worries, and many of our own. While the details must remain classified, Postol was right on some points and wrong on others. Meanwhile, none of this was lost on the Russians, who periodically and strategically commented that they understood the entire topic, and then repeatedly demonstrated that they did. We figured that less than 1 percent of the people working in missile defense understood what the Russians were saying and how important it was. We got the message.

From all of that experience, we know that the current state-of-the-art in infrared sensor technology simply won’t suffice to meet the real-world requirements, let alone the 1980s hangar queen technology represented in the antiquated STSS satellites, even with Raytheon’s late-delivered sensor upgrades that delayed the program for years and cost dearly. I tried to explain the criticality of this in a meeting with Wes Bush (now CEO of Northrop Grumman) and his STSS program manager, but it seems to have fallen on deaf ears. As soon as our teams were combined by the Defense Department and Northrop/TRW got full control of the program, they defunded work on the competing simulation as fast as possible. The reader can guess why.

In addition to the above, STSS then fell victim to poor prime contractor management and execution, with delivery three years late and ridiculous costs to refurbish existing “heritage” hardware. Northrop Grumman gave the government “heritage” all right; it was more like “Your Father’s Oldsmobile” — a 1979 model that barely runs.

Finally, the most underreported space story of the past two years was the miserable performance of STSS in its first year on orbit. Insiders know the truth about tumbling STSS satellites and dysfunctional subsystems that have barely seen the light of day. MDA knows that the Air Force bought two turkeys, and it wisely chose to quietly kill this dog.

Why doesn’t Loren Thompson ask why it has taken nearly two years on orbit to “check out” two simple STSS satellites in low Earth orbit? Formal congressional oversight questions about the awful STSS early orbit performance might be more in order.

Given the present and building U.S. budget crisis, it is highly unlikely that any STSS-type constellation will ever be deployed in low Earth orbit. Not to fear, the billions of dollars worth of U.S. missile radars deployed on land and sea around the world are plenty adequate for the Pakistani and North Korean missile threats. Just don’t mess with the Russians.

MDA made one of the smartest decisions in its history by killing STSS.


W. David Thompson, former president and chief executive officer of Spectrum Astro Inc., is now CEO of an alternative energy company that he founded. He is a fellow of the American Astronautical Society.