Because at its core Dipforge is Distribute it is inherit in the system that the Graph store is distributed. What this means is that any changes made to the primary store will be replicated automatically not only to the secondary stores in a single Dipforge instance but to the secondary stores in Secondary Dipforge instances. This means that it is very easy to setup a horizontally scaling Dipforge implementation on commodity hardware.
The Workflow engine can also scale out horizontally distributing requests to different workflow instances. This means that scaling that load out is very easy and that there will not be a bottle neck with in any implementation.
Dipforge Distributed Core
Dipforge is designed to be distributed at its core. This was designed into the following core infrastructure:
- The JNDI is designed to bind to a parent and have children bind to it. This means JNDI lookups can recurs through a tree of JNDI bindings.
- Dipforge implements a Service broker architecture. This means a Daemon or groovy end point Service can register with the Service broker and lookups for these services can be performed across the distributed tree.
- The message service in Dipforge takes full advantage of both the JNDI and Service broker architecture and can deliver a message to an JNDI bound object anywhere within a distributed architecture. The message service also provides both tradional interface defined message deliver and RPC based asynchronis messages.