The following section provides examples of various ORS 8.1.3 and later deployment models and provides an overview of behavior, for single site, single site with redundancy, multi-site, and a single Cassandra cluster across sites:
Single-Site Deployment Model
The single-site deployment example assumes that all components are planned to reside within a single box. This deployment does not provide high availability, redundancy, or failure handling. The figure below shows a single site deployment example.
Single-Site with Redundancy Deployment Model
The single site architecture with software redundancy across multiple boxes represents a standard deployment for enterprises which do not have the need or ability to provide geo-redundancy. In this deployment, two or more instances of ORS exist in a single cluster, on the same LAN.
The figure below shows a single site with redundancy deployment example.
Multiple-Site Deployment Model
In a logically separated environment such as this with multiple T-Servers, each distinct T-Server is connected to its own ORS cluster. Note that in this environment, each ORS cluster has its own private Cassandra instance.
The figure below shows a multiple-site deployment example.
Single Cassandra Cluster across Multiple ORS Data Centers Deployment Model
The figure below shows a Single Cassandra Cluster across Multiple-site Deployment example.
Single-Site with Redundancy and No Persistence Deployment Model
The following architecture with software redundancy across multiple boxes represents a deployment without Cassandra persistence/recovery.
The figure below shows a sample deployment with redundancy and without persistence to Cassandra.