Tenancy
Streamtime supports three types of tenancy models for both Kubernetes fleets and Kafka clusters. Tenancy determines the level of resource isolation and sharing between different users or workloads.
Tenancy Types
| Tenancy Type | Description | Details |
|---|---|---|
| Shared | Multiple workloads share the same Kubernetes and Kafka cluster resources. | This model maximizes resource utilization and cost efficiency, but offers the least isolation. |
| Isolated | Each workload has a logically isolated Kafka cluster within a shared Kubernetes fleet. | Kubernetes resources are shared, but Kafka clusters are separated at the logical level for improved security and performance. |
| Dedicated | Each workload has a dedicated Kubernetes cluster and a dedicated Kafka cluster. | This provides complete resource isolation, highest security, and performance guarantees. |
The tenancy model can be applied at both the Kubernetes fleet level and the Kafka cluster level. For example, a dedicated Kafka cluster will run on a dedicated Kubernetes cluster, while a shared Kafka cluster will run on a shared Kubernetes cluster.
How Tenancy Works
- Shared Tenancy: Multiple Kafka clusters run on the same Kubernetes cluster, sharing resources like CPU, memory, and storage. This is cost-effective but may lead to resource contention.
- Isolated Tenancy: Each Kafka cluster runs in its own namespace within a shared Kubernetes cluster, providing logical separation. This allows for better security and performance while still sharing the underlying Kubernetes resources.
- Dedicated Tenancy: Each Kafka cluster runs on its own dedicated Kubernetes cluster, ensuring full resource isolation. This is ideal for high-security or high-performance workloads, as it guarantees that resources are not shared with other workloads.
When to Use Each Tenancy Type
- Shared: Use when cost savings are a priority and workloads do not require strict isolation.
- Isolated: Use when moderate isolation is needed, such as for production workloads that require logical separation but can share Kubernetes resources.
- Dedicated: Use for workloads that require the highest level of security, performance, and resource isolation, such as regulated industries or high-performance applications.
Use Cases
- Shared: Development, testing, or low-risk production.
- Isolated: Production workloads with moderate isolation needs.
- Dedicated: Regulated industries, high-security, or high-performance requirements.