Offloading container storage, networking, and engines with the QingTian architecture improves resource utilization by 20% to 30%.
Millions of nodes are centrally managed to provision compute capacity to applications anytime, anywhere.
Passthrough networking through trunk ports (for VMs) and ENIs (for BMSs) delivers 40% better performance with zero loss.
The passthrough mode increases container density (up to 256 containers) on a single node.
Intelligent and hybrid scheduling of multiple service types is complemented with application topology awareness and time-based resource scheduling.
Caching and a shared view for multiple schedulers speed scheduling up to 10,000 pods per second.
Security containers are deployed on bare metal servers to provide VM-level resources and service isolation.
Container network access control and QoS rate limiting are tailored to VPC security groups.