Engineers drive fixes to Voyager 1 spacecraft and CrowdStrike software
By Steven Lincoln|December 2024
The Software Systems Technical Committee focuses on software engineering issues for complex and critical systems, including requirements, design, code, test, evaluation, operation and maintenance.
In April, the Voyager 1 spacecraft resumed transmission of science and engineering information, two days after mission controllers at NASA’s Jet Propulsion Laboratory in California sent a software fix to the spacecraft. JPL announced in November 2023 that Voyager, launched in 1977 and now traveling beyond our solar system, stopped sending “readable” data. Engineers later discovered the cause: a stuck bit in a chip that stores part of the memory of the Flight Data Subsystem, one of the spacecraft’s three onboard computers.
The software fix was created by a tiger team at JPL that systematically evaluated each subsystem on the spacecraft via a dependency bone chart until the failure was identified, a task made more difficult by the fact that “almost every engineer and programmer who helped build Voyager have either retired or passed away,” the Pasadena Star-News reported in March.
“So, a lot of what the tiger team has had to do is go back and look through old documentation,” Suzanne Dodd, Voyager project manager, told the newspaper. “Try to recreate how the code was done and why the code was done that way.”
Their solution was to divide the code into sections and store them in different places within the Flight Data Subsystem. The code sections were then adjusted so they worked seamlessly as a whole.
In July, one of the biggest unintentional software mishaps in history occurred, when a CrowdStrike Falcon Sensor security software update caused global system crashes. The outages were due to a defect in the Rapid Response Content, which went undetected during validation checks. When the content was loaded by the Falcon sensor, this caused an out-of-bounds memory read, leading to Windows crashes and the so-called blue screen of death. CrowdStrike released an update the same day as the outage that resolved the issue, and Microsoft provided instructions on how to remediate the problem ahead of the deployed fix. CrowdStrike published a root cause analysis in August that cited validation techniques, logic and valid data checks, and the company’s staged deployment approach as contributors to the crash.
Decades ago, software was considered an add-on to aerospace hardware systems. Today, however, software is recognized as an essential mission enabler that controls ground- and space-based operations of extraordinary complexity. This vital role escalates the impact of software failures. The CrowdStrike failure reportedly cost Delta Air Lines alone some $500 million in lost revenue, with its impact on other aerospace organizations still being assessed as of October. Such economic and reputational losses place increased emphasis on capabilities to verify correct functioning of critical software.
In May, the International Organization for Standardization and the International Electrotechnical Commission subcommittee on System and Software Engineering established an ad hoc group to evaluate the need for international standards on low-code/no-code software development at its biannual plenary in Berlin. The industry-specific functionalities of low-code application platforms allow users to create custom solutions for their unique needs. “The global low-code/no-code market is expected to see a compound annual growth rate of 22.7% from now until 2027,” according to a March blog post by BP3 Global, an Austin consulting firm specializing in intelligent automation. “At that time, it’s projected to be worth $86.9 billion.” Also, low-code tools are estimated to “account for over 70%” of all software development by 2025.
Artificial intelligence continued to be broadly explored for aerospace operations, but questions remain regarding whether AI providers can use copyrighted material to train their products. Among the pending lawsuits on this issue is one by the New York Times, which in late 2023 sued Microsoft and OpenAI for using millions of Times articles to train ChatGPT and other AI-based products.
