Selecting a topology
The IBM® Cloud Manager with OpenStack solution provides some predefined topologies.
The following topologies are the supported configurations for IBM Cloud
Manager with OpenStack.
| Topology | Description |
|---|---|
| Minimal | For product evaluation purposes. This topology is the simplest topology and does not require any customization. Some basic customization is supported for the KVM/QEMU compute hypervisor type. |
| Controller +n compute | For smaller test or production environments . This topology provides a single controller node, plus any number of compute nodes. You can configure this topology for your specific needs. For example, configure networking, resource scheduler, and other advanced customization. |
| HA controller +n compute | For larger test and production environments requiring highly available (HA) cloud controllers. This topology provides multiple HA controller nodes, plus any number of compute nodes. You can configure this topology for your specific needs. |
| HA dedicated nodes for Keystone and Horizon | For larger test and production environments requiring highly available (HA) cloud controller nodes and highly available (HA) dedicated nodes for Keystone and Horizon. You can configure this topology for your specific needs. |
| Distributed database | For larger test or production environments. This topology is similar to the controller +n compute topology; however, it allows the IBM Cloud Manager with OpenStack database service to run on a separate node. It also supports advanced customization. |
| Multi-region | This topology is for larger test or production environments and can include multiple hypervisor environments. This topology is similar to the controller +n compute topology; however, you can separate hypervisors by region. Each region has its own controller, but shares the same Keystone and potentially IBM Cloud Manager - Dashboard. |
The following terminology is used by the topologies:
- Deployment server
- The deployment server is the system where you install the IBM Cloud Manager with OpenStack product. The deployment server runs the Chef server and maintains the IBM Cloud Manager with OpenStack repository.
- IBM Cloud Manager with OpenStack repository
- The repository is the location that where all of the IBM Cloud Manager with OpenStack components are stored. The repository includes Chef resources, RPMs, commands, licenses, dependencies, and more. As an administrator, you manage your topology by using this repository.
- Node
- A node is a system that is part of your topology. The applicable IBM Cloud Manager with OpenStack components are deployed to the node based on your cloud environment.
- Region
- In a multi-region environment, a region is a way for you to define separate cloud deployment regions (separated by hypervisor for example), while still sharing some OpenStack infrastructure, such as Keystone.
The repository, along with the associated OpenStack Chef cookbooks and dependencies provide the basic building blocks that you can use to deploy IBM Cloud Manager with OpenStack.
The topologies support the following operating systems, database,
message queue, and OpenStack components.
Note: The following support information applies to the minimal,
controller +n compute, HA controller +n compute,
and distributed database topologies only. Also, note that not all
of the support information applies to all of these topologies previously
listed.
- OpenStack Components: Identity, Image, Network, Compute, Orchestration, Block Storage, Telemetry, and Dashboard
- OpenStack Networking: Neutron
- OpenStack Compute Scheduler: Compute scheduler filters, IBM Platform Resource Scheduler (PRS)
- Operating System for OpenStack controller: Red Hat Enterprise Linux 7.1, 7.2, 7.3, 7.4, 7.5, or 7.6
- Database: IBM DB2® (default) , MySQL, MariaDB
- Message Queue: RabbitMQ (default), Qpid
- Hypervisor types: Linux Kernel-based Virtual Machine (KVM), PowerKVM, QEMU, z/VM®, and Hyper-V
- Virtualization Manager: PowerVC
- IBM Cloud Manager - Self Service, enabled by default.