This session demonstrates a comprehensive cross-cloud expansion & intelligent massive data-flow circulation solution for the complicated cloud native AI/big data APPs scenario of an leading ISP to increase the flexibility, scalability and cost efficiency, also eliminate the data access process complexity and the low data copy efficiency.

By adopting Karmada multi-cloud orchestration tech., we are able to deploy AI data APPs cross clouds (private and public) & cross regions with unified application platform and operation/maintenance management.

The Fluid cloud-native data management tech. is used to provide a layer of efficient data abstraction (Dataset CRD) for AI data APPs, abstracting data sets into K8s standard PVC/PV resources, and seamlessly accessed by APPs through mounting operation. Then use data affinity scheduling and distributed cache engine acceleration to realize the fusion between data and computing, thereby accelerating and facilitating computing to storage access.