Defining the Platform Terminology

The BI platform includes collections of servers running on one or more hosts.

Small installations (such as test systems) can use a single host for a Web application server, database server, and all BI platform servers.

Medium and large installations can have servers running on several hosts. For example, a Web application server host can be used with a BI platform server host. This combination frees up resources on the BI platform server host, allowing it to process more information than if it also hosted the Web application server.

Large installations can have several BI platform hosts working together in a cluster.

Administrators need to be familiar with the following BI platform components and terminology.

Host

This term is used for the physical or virtual machine (PC tower, server blade) that is hosting nodes. Every single machine will have at least one node/Server Intelligence Agent (SIA). More than one node could be created if needed.

Node

A node is a collection of BI platform servers running on the same host. One or more nodes can be contained on a single host.

Service

A service is a server subsystem that does a specific function. The service runs within the memory space of its server under the process identifier (PID) of the parent container (server).

Server

A server is an operating system level process hosting one or more services, for example, the Adaptive Job Server. A server runs under a specific operating system account and has its own PID.

Scaling a system

Server processes can be vertically scaled, where one computer runs two or more servers (such as the Adaptive Job Server). Server processes can scale horizontally, where server processes are distributed between two or more networked computers.

The advantages of having several servers include:

• Improved performance

  Several BI platform server hosts can process a queue of scheduled reports faster than a single BI platform server host.

• Load balancing

  If a server is experiencing a heavy load, the CMS automatically sends new work to other servers in the cluster.

• Improved availability

  If a server comes across an unexpected condition, the CMS automatically re-routes work to different servers until the condition is corrected.