|
Scalability
The weblicon PIM system architecture is designed to provide a highest level of scalability offering near-linear scalability at all relevant tiers. The architecture can be scaled from simple installations serving a few thousand users to multi-million user installations handling hundred thousand or more concurrent user sessions.
Scalability at the HTTP server and application tier can be implemented by the WebObjects application server architecture, which is proven to handle very large installations. It can handle an arbitrary number of clients connecting via multiple HTTP servers to multiple application servers. Each HTTP server has a WebObjects HTTP adaptor installed dispatching requests to multiple application servers running multiple application instances. Each application instance can handle multiple sessions supporting multithreading and symmetric multiprocessing. Individual sessions are handled in separate threads ensuring a balanced performance among the different sessions.
HTTP Server Tier
The HTTP server tier can be scaled linearly by installing additional HTTP servers for handling incoming requests from HTTP clients as the load increases. The HTTP servers are running identical configurations and can be included in the load-balancing mechanism without interruption of the service. Incoming requests from HTTP clients are dispatched via the WebObjects HTTP adaptor to the application server tier.
Application Server Tier
The application server tier can be scaled linearly by installing additional application servers running identical configurations as the load increases. The application servers are seamlessly included into the load-balancing of the WebObjects HTTP adaptor without interruption of the service. This architecture provides near linear scalability of the application server tier from several hundred to hundreds of thousand concurrent user sessions.
Database Server Tier
Clustering architecture of leading RDBMS vendors such as Oracle provides scalability to the database tier. Additional database servers can be added to the cluster without reconfiguring the application servers or manual repartitioning of the database. Similar approaches can be used to scale the LDAP servers by using a combination of master/slave and dual-master configurations as well as partitioning of the LDAP tree. Using clustering and partitioning, the database tier can be scaled to handle the data and traffic of multi-million user installations. Access to the file server can be scaled by load balancing the WebDAV requests to a cluster of file servers handling the actual storage.
|
  |
|
|
