Hybrid Data Pipeline can be deployed on a standalone machine or on one or more nodes behind a load balancer. Many configurations and best practices are contingent on how Hybrid Data Pipeline has been deployed.
For a production environment, Hybrid Data Pipeline should be deployed on one or more nodes behind a load balancer to support scalability and availability. In a load balancer deployment, client application requests must be directed to the load balancer which forwards requests to the node or nodes running the service. When multiple nodes have been deployed, requests are distributed across the cluster. See Load balancer deployment for more information.
When deployed on a standalone node, the service is installed on a single host machine that manages all queries, simplifying maintenance and administration. A standalone deployment is not recommended for a production environment because it does not provide the scalability and availability of a load balancer deployment. However, a standalone deployment may be required due to resource limitations and other restrictions. If a standalone deployment is required in production, then, as a matter of best practices, the deployment should include an external system database and a user-specified key location. See Standalone deployment for details.
Important: There is currently no migration path from a standalone deployment to a load balancer deployment. Therefore, a standalone deployment is not recommended for environments where scaling up the service may be desired. A standalone node deployment is also not recommended for security and system recovery purposes. If you want to move from a test environment to a production environment, you should begin by deploying Hybrid Data Pipeline on a single node behind a load balancer. When deploying the service on a single node behind a load balancer, you can increase availability and scalability as demanded, and address security and recovery concerns as required.
Whether you deploy the service on a standalone node or behind a load balancer, Hybrid Data Pipeline can be run on-premises or in the cloud. See the following topics for more information.
In addition, after at least one installation of the Hybrid Data Pipeline server, you can modify your environment to use an external JRE at runtime as opposed to the embedded JRE that is shipped with the product package. See External JRE support and integration.
