Concept

In the following it is assumed that the FirstSpirit server services are to run on different systems and a cluster solution is to be used in the area of the application server.

The architecture shown in the following figure then results:

Figure: High-availability cluster with load distribution

The FirstSpirit Client (both the SiteArchitect and the ContentCreator) communicates with the clustered web application server via HTTP(S). The FirstSpirit sessions are distributed to the cluster's individual application servers via the HTTP(S) Load Balancer. Load distribution can take place on URL basis and/or on HTTP Session basis.

Behind the application server cluster, a range of FirstSpirit services are started on different systems:

FirstSpirit Master server: The FirstSpirit Master server centrally manages all FirstSpirit projects and deals with the users' queries/changes and distributes the tasks, wherever possible, to other FirstSpirit servers.

FirstSpirit Generation servers: Several FirstSpirit generation servers should be used for complex or frequent deployment processes to move the load from the master when generating the web presence. If necessary, several deployments can also be distributed to different servers in this way. A generation server contains an RORS.

ReadOnlyRepository server (RORS): A special Repository Manager processes the queries from a Generation server.

The individual FirstSpirit servers decide at the start which schedule area they are responsible for and which managers have to be started for this. A FirstSpirit generation service is created, for example, from a “normal” FirstSpirit server, which only starts the servers and generation manager required for communication. The FirstSpirit software architecture is set up so that (in simplified terms) a generation manager is used when a schedule is executed whereby it is not possible to tell whether it runs locally or remotely.