Distributed computing
Distributed computing involves the cooperation of two or more machines communicating over a network. The machines participating in the system can range from personal computers to super computers; the network can connect machines in one building or on different continents.
- Share the cost of expensive resources, such as a typesetting and printing service, across many desktops. It also gives you the flexibility to change the desktop-to-server ratio, depending on the demand for the service.
- Allocate an application’s presentation, business, and data logic
appropriately. Often, the desktop is the best place to perform the
presentation logic, as it is nearest to the user and can provide
highly responsive processing for such actions as drag and drop GUI
interfaces.
Conversely, you may feel that the best place for the database access logic is close to the actual storage device - that is, on an enterprise or departmental server. The most appropriate place for the business logic may be less clear, but there is much to be said for placing this too in the same node as the data logic, thus allowing a single desktop request to initiate a substantial piece of server work without intervening network traffic.
Distributed computing enables you to make such trade-offs in a flexible way.
Along with the advantages of distributed computing come new challenges. Examples include keeping multiple copies of data consistent, keeping clocks in individual machines synchronized, and providing network-wide security. A system that provides distributed computing support must address these new issues.
- Distributed Computing Environment (DCE)
- The remote procedure call model implemented by the Open Software Foundation's DCE is supported in CICS.
- Distributed program link (DPL)
- This is similar to a DCE remote procedure call. A CICS client program passes parameters to a remote CICS server program and waits for the server to send data in reply. Parameters and data are exchanged by means of a communications area.
- The external CICS interface (EXCI)
- An MVS™ client program links to a CICS server program. Again, this is similar to a DCE RPC.
- The external call interface (ECI)
- The ECI enables CICS Transaction Server for z/OS® server programs to be called from client programs running on a variety of operating systems. For information about CICS Clients, see the CICS Transaction Gateway Programming Guide.
- Function shipping
- The parameters for a single CICS API request are intercepted by CICS code and sent from the client system to the server. The CICS mirror transaction in the server executes the request, and returns any reply data to the client program. This can be viewed as a specialized form of remote procedure call.
- Asynchronous transaction processing
- A CICS client transaction uses the EXEC CICS START command to initiate another CICS transaction, and pass data to it. The START request can be intercepted by CICS code, and function shipped to a server system. The client transaction and started transactions execute independently. This is similar to a remote procedure call with no response data.
- Distributed transaction processing
- A program in the client system establishes a conversation with a complementary program in the server, and exchanges messages. The programs may use the APPC protocols.
- Transaction routing
- Terminals owned by one CICS system to run transactions owned by another.
The CICS family of products runs on a variety of operating systems, and provides a standard set of functions to enable members to communicate with each other.