Replies: 1 comment
-
Thanks for the interest, @abdennour! The Kamaji scope is to simplify the life cycle of a fleet of Kubernetes clusters in a hyper-converged environment. Thus, we're aiming in designing it in a non-opinionated way to ensure to make it is pluggable in any infrastructure, with limited requirements and constraints. Multi-cluster is a system design used to deploy applications in separate contexts for several reasons, and in some circumstances, multi-cluster application communication must be provided. There are plenty of solutions out there that are already trying to solve this need, both from the data plane standpoint (thinking of Cilium at the CNI level, and KubeSlice, etc.) or at the application level (Submariner, Liqo, etc.). Our actual roadmap is to focus more on the core idea, simplifying Day2 operations and reducing the operational burden, allowing any end-user to pick their preferred way to let the tenant control plane cluster applications communicate with each other. At the same time, we strongly believe in the open-source methodology, as well as an open governance model, and I'd like to thank you for opening this issue that can be a starting point to discuss with the community any feature request or demand, by also comparing to other multi-cluster (and hyper-cluster) solutions. |
Beta Was this translation helpful? Give feedback.
-
Since all clusters are provisioned with the same admin cluster, is there a networking solution to let pod from cluster tenant A talks with pod from cluster tenant B ?
Something like submarina
Beta Was this translation helpful? Give feedback.
All reactions