Scale mobile enterprise solutions as demand increase



The RAMP server has been designed to take advantage ofscalability, enterprise scalability, mobile development platform,  j2me mobile applications, both a scale-up and a scale-out approach to capacity and high availability.

Scale-up approach


The RAMP platform has been designed to run efficiently within a multi-processor hardware setup. This is critical when a scale-up approach is followed where it is entirely possible to have a single large server with anywhere between 1 - 64 CPUs. Developing parallel server applications that can take advantage of multi-processor servers is a highly skilled process and the RAMP server was designed to run in such an environment.

High availability is provided by the hardware vendor that provides redundancy and performance guarantees for the underlying hardware components.

Scale-out approach


The RAMP mobile enterprise application platform has been designed to run in a scale-out approach. The RAMP server clusters the various application server nodes in an active-active configuration so that when the load increases it is simply a case of adding additional server nodes to the cluster to increase capacity. The RAMP server scale-out implementation can scale to dozens of machines. The deployed system can be illustrated in the picture.

The load balancing servers are the first point of entry to the deployed system for a connecting client. They are responsible for delegating requests to the application servers. The load balancers will monitor the load on the application server nodes and if one or more nodes are under too high a load they will redirect work to nodes that are under a lesser work load. The load balancers can also detect if an application node becomes offline and redirect work to other nodes.

The state clustering servers (Terracotta servers) ensure that the application servers have consistent access to client session information so that should one application server fail, another application server can seamlessly take over its work load.


Other information on RAMP: