This video explains and demonstrates simplyblock’s hyperconverged deployment mode, where storage and compute run on the same Kubernetes nodes.
In hyperconverged mode, simplyblock uses the local NVMe drives on each node as storage while still providing disaggregated access: any pod on any node can access any volume, not just the volume on the same physical node where the data lives. This delivers the hardware efficiency of hyperconverged while keeping the flexibility of shared-everything storage.
Topics covered:
- When to choose hyperconverged vs. fully disaggregated simplyblock deployment
- How NVMe/TCP enables cross-node volume access even in hyperconverged mode
- Erasure coding across nodes to maintain data protection without full replication overhead
- I/O path from the application pod through the CSI driver to the local NVMe device
- Failover behavior when a node goes offline in hyperconverged topology
The hyperconverged deployment is well-suited for bare-metal Kubernetes clusters and edge deployments where separate storage nodes are not practical.
See the simplyblock deployment modes page for a comparison of hyperconverged vs. disaggregated architectures.