Multi-Purpose Cloud Filesystem
summary
SaaS DR replaces a second data center by streaming incremental backups into a cloud filesystem that launches VMware clusters on demand; steady state stores deduplicated content‑addressable clumps (1–64KB) packed into 8MB LFS objects on S3, using EC2 + local NVMe read caches for instant high performance. Metadata lives in LSM trees to support millions of snapshots and clones, snapshots are immutable and hidden for ransomware protection, and the service claims low steady‑state cost with fast recovery and built‑in integrity checks.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:28Disaster recovery importance
- 3:54Traditional DR problems
- 6:42Simplifying DR with SAS
- 10:11Designing the filesystem
- 12:36Layered architecture
- 15:05Ensuring data integrity
- 17:56Selling the solution
- 21:25Conclusion and key takeaways
related talks
09mShared cloud file storage, uses cases, and the latest Amazon EFS innovations
29mFuture-proof your file-based workloads with FSx for Windows File Server
30mWhat's New in File Storage
56mChoosing the right storage service for the job
18mKeynote: Accelerate business agility with data in the cloud
28mCloud The Final Destination
20mRadically simple data protection with AWS and Druva
37mCloud, security, and science converge in a data-driven world (PRT324)