Data Distribution Connectors
Data distribution connectors can involve a composition of different connector types:
- Procedure call
- Event
- Arbitrator
- Data access
- Stream
- Distributor
A data distribution connector can combine these connector types to support the movement and distribution of data between components.
Event-Based Data Distribution
Event-based data distribution connectors are composed of:
- Event
- Data access
- Stream
- Distributor
They send and receive data through asynchronous notifications (events).
An example is publish-subscribe middleware.
Client-Server Based Distribution
Client-server based distribution connectors allow seamless distribution of data between distributed systems using RPC (Remote Procedure Call).
They are composed of:
- Procedure call
- Data access
- Stream
- Distributor
Invocation occurs through a synchronous remote procedure call that appears as if it were a local method call.
Examples include:
- HTTP/REST
- Java RMI
- CORBA
- FTP
- SOAP
P2P Based Data Distribution
P2P-based data distribution connectors are composed of:
- Arbitrator
- Data access
- Stream
- Distributor
They:
- Use arbitration for synchronization and invocation.
- Use rendezvous to achieve concurrency and scheduling.
- Often provide transaction support.
- Can roll back an invocation when necessary.
An example is BitTorrent.
