This page describes the key platform components and their installation on different appliances in production and test environments.
Abiquo platform servers
The Abiquo UI lets cloud consumers interact with the resource cloud through the Abiquo platform servers. You can install each of the following components on one or more appliances in a production environment or on fewer appliances in a test environment.
Platform server | Description |
|---|---|
Server | The Abiquo Server contains the API and the UI. And it contains the business logic and it manages all the private datacenters and public cloud regions. It also handles and stores all of the information and events created in the cloud environment, using the database on the Datanode server or, in a test environment, on this server. |
Remote Services | The Remote Services server performs operations requested by the Abiquo Server on the infrastructure, such as deploying a VM, or duplicating a VM template. Each private datacenter has one Remote Services server and public cloud regions can share Remote Services with private cloud datacenters. For example, when a cloud user creates a volume in public cloud, Abiquo communicates with the discovery manager, which communicates with the API of the public cloud provider to create the volume. |
V2V Services | The V2V (Virtual to Virtual conversion) services server has one of the remote services (BPM) but you will usually install it on a separate server for performance reasons. The process runs when users download VM templates, create VM instances, or when they upload new templates, to convert them to hypervisor formats. |
Datanode | The Datanode server contains the Abiquo database and services, such as RabbitMQ for queues, Redis for datacenter subscriptions, and Zookeeper (for API load balancing). These components may also be installed on the Abiquo Server in a test environment. |
Monitoring | The Monitoring Server contains the Abiquo Watchtower monitoring services (Emmett, Delorean, KairosDB and Cassandra) for the platform. You can deploy monitoring as a standalone service or as a cluster (for large environments). |
Catalogue
A key element in private cloud is the catalogue. In private cloud, the cloud and tenant administrators upload VM templates to the catalogue, for users to easily create their own VMs. The catalogue stores disks on NFS shares called NFS repositories. Each private cloud datacenter must have its own private NFS repository. In public cloud regions, Abiquo caches the details of the cloud provider templates but it does not store the disks, so you do not require an NFS repository.
The diagram below shows an Abiquo Server managing several private datacenters in a distributed environment. The main point to note is that each private cloud datacenter has its own NFS repository and Remote Services server.
Abiquo installation profiles
Abiquo components are modular and it is possible to install them on multiple servers. Abiquo has the following main installation profiles:
Monolithic with all the components on a single appliance, for test environments.
Distributed scalable with each component on its own appliance
Distributed HA cluster with multiple load balanced API/UI nodes, datanodes, and monitoring nodes.
Test installations
For proof of concept and other testing, you can use a Monolithic environment that has all of the Abiquo components on the same appliance. A test environment can also include a monitoring appliance as in a production installation (although this is not shown in the diagram below).
For all the requirements to install a monolithic test environment, see Monolithic environment and for instructions, see Install a test environment.
Production installations
In production installations, the components are configured as Distributed environments. This means that each server has separate Abiquo components for improved performance and reliability.
Distributed environments require at least one Server, and at least one datanode (database and services), then for each datacenter, they will require Remote Services and V2V Services. Most environments will also have at least one monitoring appliance.
For high availability, in an HA cluster environment the API and UI nodes, the Datanodes, and the Monitoring nodes are usually configured as clusters, for example, with 3 servers each and floating IPs.
The following diagram shows a distributed environment.
In production environments, if you are using public cloud only, the platform does not require the V2V Services or the NFS repository.
For more information, see Distributed HA cluster environment.
Related pages