Thales updates entertainment content without grounding flights

THE CHALLENGE

Keeping content fresh without grounding aircraft

For long-haul travelers, in-flight entertainment (IFE) is a key part of the travel experience, helping passengers pass the time during long flights. Managing IFE updates and delivery is a responsibility that French multinational aerospace corporation Thales has handled for over 30 airlines.

Keeping content fresh and easily accessible is a core task that previously required a technician to physically visit each aircraft once a month, keeping flights on the ground and potentially impacting airline revenue.

“There were many limitations with the older systems,” said Alexandre Plaitant, DevOps Engineer at Thales. “We wanted to give airlines the autonomy to deploy the applications they want on board, and update content incrementally—potentially daily—without the need for physical intervention.”

The Thales system uses two main database environments: one on each aircraft, and a parent database on the ground that handles content distribution and updates. While each aircraft uses MongoDB Community Edition as its local database, Thales originally chose Azure Cosmos DB to handle ground-based activity. 

“An aircraft isn’t always connected to the internet, so we have to host all the content on board, which requires substantial storage,” added Plaitant. “Our goal is to provide the best traveler experience and be as responsive as possible. The challenge is synchronizing between the ground and the aircraft—when internet access is available—to update the content.”

Cosmos DB worked well at first, but began to struggle as the Thales application grew. Response times became excessively long, and performance bottlenecks caused stability issues. If a service launches too many requests at once, the database could crash, taking down the entire fleet management platform.

Additional issues also emerged. A lack of transparency in the previous solution’s pricing model made it difficult to scale and manage costs effectively. Cosmos DB’s API was also tied to an older version, compatible with MongoDB 4.0, causing engineering problems that prevented Thales from implementing more modern features.

“Some drivers were incomplete, which meant certain features were missing,” explained Plaitant. “That forced us to reimplement processes directly in our code—work we didn’t want to spend time on.”

“We were stuck with a limited version of Cosmos DB that was starting to create security issues and affect the overall passenger experience,” added Jeanne Barriere, Full Stack Developer at Thales. “We realized that the best way to keep everything up to date was to switch to MongoDB Atlas.”

OUR SOLUTION

Delivering a quick and straightforward migration

Thales needed a solution that offered better scalability, greater transparency, and significantly lower latency to ensure reliable operations. By choosing MongoDB Atlas to handle its ground-based infrastructure, Thales successfully resolved the significant technical debt associated with Cosmos DB and consolidated its technology stack.

“We already had MongoDB on board the aircraft, so it made sense to run MongoDB Atlas on the ground,” said Plaitant. “It also helped with our resource consumption, as MongoDB isn’t very resource-intensive.”

Thales used standard command-line utilities to move its data from the Cosmos DB instance and push it into MongoDB Atlas. It was a straightforward approach that meant Thales did not need complex, custom extract, transform, load (ETL) tools to move the data.

“We updated the MongoDB dependency, and that was that. It was very easy,” said Plaitant. “It required virtually no modification to the code itself.”

The architecture also allowed Thales to perform these migrations without losing data. It simply shut down specific microservices briefly to stop event processing, performed the database migration, and then restarted the services in MongoDB Atlas.

“Most jobs were migrated in under a minute, and services with smaller databases took just a few seconds,” added Barriere. “Everything worked perfectly. It was easy and cost very little.”

OUTCOME

A tenfold reduction in key response times

The migration from Cosmos DB to MongoDB Atlas delivered key business, technical, and operational outcomes. It gave Thales better visibility into pricing, enabling it to reduce overall costs while delivering a more reliable, consistent service.

It also provided access to modern features unavailable in Cosmos DB and resolved issues such as missing native aggregations and incomplete drivers. Instead of reimplementing standard database functionality directly in their application code—work that added unnecessary complexity—developers now have the freedom to innovate.

“Steps like that are now handled automatically by MongoDB Atlas,” said Barriere. “Frankly, the migration has been a breeze.”

The most significant shift is that airlines can now manage and update content independently. This eliminates the need for maintenance technicians to board each aircraft with a hard drive once a month.

“Thanks to MongoDB Atlas, airlines can run daily incremental updates of their content and manage it themselves,” said Barriere. “We can also store information in the onboard database and retrieve it on the ground at the end of each flight. That gives us valuable remote monitoring and tracking of every aircraft.”

Switching to MongoDB Atlas also resulted in a significant performance boost, with response times for some processes now reduced by a factor of 10. A previously vulnerable system is now resilient and available, with benefits for both engineers and developers.

More importantly, passengers now enjoy a smoother, more accessible in-flight experience, with minimal latency, a responsive system, and up-to-date content. Aircraft simply query MongoDB Atlas whenever they have internet access, and download updates as they become available—and critically, without keeping aircraft grounded for hours.

“Even if an aircraft only has time to download 50% of the new content, it can pause the process and resume it when connectivity returns,” said Plaitant. “We no longer have to immobilize a plane just to deliver an update. The performance gain with MongoDB Atlas has been amazing.”