A cloud (shown in Figure 2.3) is a collection of virtual machines; each virtual machine runs a separate copy of an operating system and an installation of Postgres.Figure 2.3 - Using EDB Ark in a Cloud.You can specify different combinations of CPU speed, RAM, and disk space to suit your needs when provisioning an EDB Ark cluster.When using OpenStack as a cloud provider, an OpenStack image must be registered for use as an EDB Ark server image. Each EDB Ark server image specifies the image ID of an OpenStack image and the name of the default_user that is specified in the /etc/cloud/cloud.cfg file associated with that image. You must register the OpenStack image in the EDB Ark Administrator’s console before using it to create an EDB Ark database engine definition.After describing the server image in the EDB Ark Administrator’s console, an administrator can use the server image to define an EDB Ark database engine. A database engine is a combination of an OpenStack virtual machine and a database type.Figure 2.4 – Using an OpenStack image with EDB Ark.The Administrator can use the same server image to create multiple database engine definitions. (see Figure 2.4). For example, you can create database engines for both PostgreSQL 9.4 and EDB Postgres Advanced Server 9.5 that are both provisioned on the same underlying server image CentOS 6.5 x86_64.When a user defines a cluster, the Ark console uses the information in the EDB Ark server image to launch a virtual machine (specified by the OpenStack image) to host the database server. The end-user selects the cluster configuration (the DB Engine type, size, speed and scaling preferences) in the EDB Ark end-user console.