Server room
3 min

S3-native storage for AI predictive maintenance with Scality ARTESCA - automotive case study

MIchael Anderle
Michael Anderle

Executive summary

For the "Magic Eye" predictive-maintenance project, Eywo built an S3-native object-storage layer on Scality ARTESCA that cleanly separates image capture from AI compute letting storage and AI scale independently on standards-based interfaces.

About client

One of the world's oldest active automakers, a wholly owned subsidiary of a major global automotive group, delivering over a million vehicles a year. The manufacturer is executing two strategic shifts, transitioning to dedicated battery-electric platforms, and expanding localized production and distribution into high-growth Indian and Southeast Asian markets.

The challenge

Magic Eye places cameras on production equipment to catch early signs of wear, feeding the images to an AI model for predictive maintenance. Before the AI can analyse anything, those images need somewhere reliable to live. Raw storage speed was not the priority what mattered was full, native S3 protocol support, so the AI platform (running in its own separate compute environment) could pull images in a standardized, predictable way. The architecture needed a clean split. cameras write on one side, AI reads on the other, reliably and without friction.

The solution

Eywo designed the architecture around Scality ARTESCA as the central object-storage layer, built for image-heavy, S3-native workloads. During implementation, ARTESCA was deployed and integrated into the client's infrastructure. The critical phase was proving the full round trip, cameras securely writing images to ARTESCA, and the AI platform reliably pulling that data back via the S3 API, without gaps or bottlenecks. Once validated in test, the solution moved to production and into Eywo's ongoing operational support.

Lessons learned

  • Separate storage from compute. The highest-performance storage system wasn't what this use case needed.
  • Standards-based interoperability was the real success factor solid S3 support let storage and AI compute scale independently.
  • ARTESCA proved a stable, accessible, easy-to-integrate hub, keeping storage simple and standards-compliant made the whole predictive-maintenance pipeline easier to build and run.

Eywo designs and implements observability, ITSM, and infrastructure solutions for enterprise clients across regulated industries. If you're planning a similar project, get in touch.


Let's talk