Business Objects  

Distributed Architecture

A distributed architecture is one in which common server resources are shared among the server products, and are distributed over several physical servers. WEBINTELLIGENCE modules, or processes, can run on different machines, and the system automatically recognizes all the machines in the WEBINTELLIGENCE system, as well as the distributed services at runtime.

You can therefore scale up the WEBINTELLIGENCE system as demand grows by adding more servers to the system to share the transaction load and provide failover capabilities. This way, if one server fails, the system automatically redirects requests to another, functioning server in the system.

This type of distributed WEBINTELLIGENCE structure is called a cluster.

Do you have to use a distributed architecture?
No. 

You can run WEBINTELLIGENCE and all its components on a single server if your deployment is small and simple enough.
However, using a distributed architecture deployment allows for much larger deployments than a traditional client-server architecture can handle. Distributing components over several servers optimizes use of the different server resources, reduces the workload on the web server, and increases performance. It can also provide failure recovery if the components are installed on more than one server,
as well as help isolate the processes in case of failure.

A distributed architecture deployment including multiple Business Objects server products reduces the total cost of ownership for large-scale deployments.

• It provides self-service data access for end users, and high availability and performance on a full 3-tier or n-tier architecture.

• It allows you to cut maintenance and deployment costs by using zero administration clients (only the application servers to maintain).

• It uses Java applets, which ensures platform independence and low cost of ownership as well as maintaining the intuitive drag and drop interface that full client users are used to. The compact Java applets are automatically downloaded to the thin client web browser.

• All processing is carried out on the middle-tier application server.

All of these advantages are made possible by the use of a CORBA architecture.