Objective: "Speedup data access for diversified scenarios on cloud"
- Support dataset metadata backup and restore (for small files)
- Support more cache runtimes (eg. Vineyard) for diversified data types
- Optimize cache runtime for the running apps on-the-fly
Objective: "Make cache runtime and application scheduling intelligent"
- Flexible scale in/out capability of cache runtimes (HPA)
- Workload-specific data cache orchestration
- Intelligent Data Prefetch based on Workload Scheduling History
Objective: "Enable operation and maintenance of Fluid with less cost"
- Data Abstraction: More data abstraction types for different computing frameworks
- Observability: Enhance observability for cache runtimes
- More Operations: Enhance Dataset Operation (Pin/Load/Free..)
- Data Management: Dataset Cache Garbage Collection in Kubernetes