DEVOURED
Cloud Native Computing Foundation Announces Karmada Graduation
Karmada has officially graduated from the CNCF, providing a production-ready standard for orchestrating applications across multiple Kubernetes clusters and hybrid clouds.
What: Karmada v1.19 has achieved CNCF graduation, marking its maturity for large-scale enterprise use by companies like Bloomberg and Alibaba. The project provides cross-cluster scheduling, failover, and multi-region application management without requiring changes to existing Kubernetes resource definitions.
Why it matters: Standardizing multi-cluster orchestration is becoming a requirement for enterprises dealing with AI training and inference at scale, as they must move workloads across disparate GPU-enabled zones and geographic regions.
Deep dive
- Karmada extends the standard Kubernetes API to support cross-cluster placement, propagation, and failover.
- The v1.19 release enhances priority-based scheduling and multi-component coordination for distributed AI training workloads.
- The project underwent a third-party security audit and established a formal steering committee as part of its graduation process.
- Supported adopters include Bloomberg, Trip.com, and Alibaba Cloud, who use the tool for hybrid cloud resilience and resource pooling.
- Roadmap items for 2026 focus on resource-aware control planes and dynamic resource allocation for GPU accelerators.
Decoder
- Multi-cluster Orchestration: The management of containerized workloads spanning two or more independent Kubernetes clusters, often across different cloud providers or regions.
Original article
Full article content is not available for inline reading.
Explore all 55 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi