Bartłomiej Święcki

Immutability is a powerful concept. Without it we wouldn’t be able to rely on bank transactions, medical records or even some of our private data. And we usually take for granted that our data is safe somewhere out there in the "cloud". But is it? There are well-known multi-billion dollar stories where attackers were able to break into well-secured systems and modify data stored in databases.

What if there was a way to protect the immutability with math? What if any attempt of historical data manipulation could easily be detected? In this talk I’ll share how we do it in immudb - a new database built on a zero-trust model.

In this talk I would like to share my experience with large-scale backups of RBD images in Ceph clusters at OVH. I will talk about challenges that we have faced when developing such solution and provide some practical guidelines for everyone willing to implement similar solution.

This talk is intended for everyone who would like to be extra safe with their software defined storage.

• Why backup Ceph if it's already replicated?
• Estimating resources needed for backup
• Challenges of implementing backup at large scale
• Hot/cold backup strategy
• Some backup-related numbers at OVH