Deployment Model Comparison MatrixΒΆ
This section helps you evaluate and select the most appropriate deployment model for your environment by comparing key features.
Feature |
AIO-SX |
AIO-DX |
Standard |
Distributed Cloud |
|---|---|---|---|---|
Node Count |
1 |
2 + up to 50 workers |
2 controllers + up to 8 storage nodes + 100 worker nodes |
Multi-site |
Redundancy |
No |
Yes |
Yes |
Yes |
Max Worker Nodes |
1 |
50 |
100 |
Varies by site |
Storage Type |
Rook-Ceph |
Rook-Ceph |
Rook-Ceph (Controller, Dedicated, or Open Model) |
Rook-Ceph (geo-distributed) |
Storage Scalability |
Low |
Moderate to High |
Moderate to High |
High |
HA Controller Services |
No |
Yes |
Yes |
Yes |
Ceph Replication |
Optional |
Yes |
Yes |
Yes |
Target Environment |
Small edge sites |
Small to medium edge deployments |
Medium to large enterprise |
Telco core (central sites), global enterprise |
Networking Complexity |
Low |
Moderate |
Moderate to High |
High |
Deployment Cost |
Lowest |
Low |
Moderate to High |
High |
Extensibility |
Limited |
Moderate |
High to Very High |
Very High |
Note
From a Rook perspective, there is no difference between AIO-DX+ and STD+workers deployments. Both can use any deployment model, and the Ceph cluster size depends on the number of OSDs available across the nodes. However, an AIO-DX deployment without workers is more limited because combined application and storage workloads on a single node can quickly hit resource constraints (CPU and memory).
Note
Kubernetes automatically reschedules containers from failed worker nodes to healthy ones in all HA configurations.