Remember when Europe’s entire Galileo satellite system fell over last summer? No you don’t. The official stats reveal it never happened

We're too outraged to do a Bohemian Rhapsody headline

It was a devastating blow to the credibility of Europe’s Galileo satellite project: the navigation system fell over during an upgrade in July, requiring a reboot that took six days. Now it appears it officially never happened.

Billions of organizations, individuals, phones, gizmos, apps, and so on, across the globe simply stopped listening to the system, the entire purpose of which is to provide state-of-the-art positional accuracy to within 20 centimetres yet was off by more than 500 metres.

The European Union ordered an independent inquiry, which issued several recommendations in November though has yet to explain exactly what went so badly wrong.

So it came as somewhat of a surprise when the system’s quarterly performance report [PDF], published last week, revealed that the system had surpassed all its targets and was operating optimally across July, August and September – a veritable sea of green boxes.

Pleased to report everything went well. No problems at all. Green all the way ... Well, except for the whole system going down for six days.

“During this quarterly reporting period, the measured Galileo Initial Open Service performance figures exceed the Minimum Performance Level (MPL) targets specified in the [OS-SDD], with the exception of the UTC availability MPLs in July,” the report reveals.

Yes, despite the entire system going offline and being unusable for a week, the European Global Navigation Satellite Systems Agency (GSA) is proud to report that it has hit all its targets so you can keep sending the cash.

Average is your friend

How is this even remotely possible? With high-school-level stats manipulation.

Yes, for six days in July, there may have been a few zero per cent days or moments, but average it out across the whole month and you’re looking at a pretty impressive 81.27 per cent availability, surpassing a surprisingly low 77 per cent target.

Failing that, why not average figures out across the entire quarter? “The Availability of GGTO Determination metric was 95.68 per cent over the whole quarter,” the report proudly announces… The measured values are comfortably above the [OS-SDD] MPL target of 80 per cent.” It’s almost as if THE ENTIRE SYSTEM didn’t collapse. Nothing to see here. Move along. Move along.

As for the actual event itself, whose details are still pretty much secret, that is covered in a two-page annex at the end of the 40-page report.

“On July 10th 2019, Galileo was affected by a technical incident related to its ground infrastructure, which resulted in an interruption of the Galileo initial navigation and timing services,” it notes.

“The technical issue was solely related to the ground infrastructure in the Galileo control centres, not to the Galileo satellites. The incident impacted the time and orbit determination function. It was caused by a series of unrelated events that impacted the synchronisation of elements of the Galileo ground system.” The report continues in this vein: downplaying what was, objectively, an enormous failure.

If you squint, you can see the outage here. The 77% availability target was just met! I love Galileo & the people working on it, but having a huge outage and then declaring 'all green!' because after sufficient massaging you met your 77% performance target is just 😠. pic.twitter.com/e8gEjrjHZS

— GalileoSats 🛰 (@GalileoSats) January 28, 2020

But it gets worse: the full explanation of what happened is that it was “caused by a combination of events that occurred quasi-simultaneously and independently, in a context of temporary limitation of redundancy” – a vague and deliberately worthless statement.

Mission accomplished

What about the independent report ordered by the European Commission? It is mentioned in an almost deliberately unhelpful high-level way: “The European Commission set up an independent Inquiry Board in September 2019 to analyse the root causes of the incident and provide recommendations. The Board was composed of high-level members and experts with proven track records in complex operational projects, in the transport and defence sectors.

“The Board delivered its final recommendations to the European Commission at the beginning of November, to be put into operation at programme and service provision management level. At the time of publication of this report, the Galileo programme has developed an initial action plan for the implementation of the Board’s recommendations, with several of them being already accomplished.” It doesn’t say what the recommendations are.

That painful lack of information is also present in the EC’s official announcement of the independent report. It lists the recommendations and couldn’t be vaguer if it tried:

“In particular, the Board recommends to:

• Review Galileo’s operational management to better meet the needs of a service-driven exploitation phase and parallel evolution, while ensuring service continuity, integrating an oversight function;
• Improve service continuity, system stability and system resilience, including operability;
• Enhance operation, maintenance and configuration management, including training;
• Ensure prompt and structured institutional communication towards users and Member States in crisis situations."

Nine lives

The full report has yet to be released. We have asked for a copy. But according to those in the know, it identified nine different things that should have prevented the system from falling over and yet went wrong.

It’s hard to know whether to be relieved nine different factors could have prevented the system failure; or scared that Galileo’s systems failed to kick in and prevent the outage no less than nine times.

People’s confidence in the Galileo system had already been shaken by this incident. The fact that the constellation of organizations around Galileo can’t even face up to their own failings serves as a bright red flag to the entire world that the culture around Europe's satellite project is broken.

The same color red that should have appeared all over this report but for some reason came out as green. ®