Designing Data & Process Placement

To allocate the processes and data to specific system processor nodes.

The Physical Data Model and Physical Process Model, or the Object Model, identify everything that must be placed within the distributed architecture.

Define Division of Client and Server Processing

Finalize the processing to take place on the server and the client. There are various degrees of division ranging from monolithic (all processing on the server) to fully distributed.

Re-examine the processes defined to ensure that the distribution of data and processes as defined in the data and process placement strategy remains appropriate for the processes as designed.

Define Clients and Servers

The data and process placement strategy defines the approach to distribution of data and processes.

Identify the distribution of data and processes among the different client and server nodes in the network. This could be a geographic distribution, such as branches located across a region, or a distribution by business division, for example, sales and finance.

Identify all server nodes in the system and allocate server processes and interfaces to appropriate physical processor nodes. Identify all client nodes in the system and allocate client processes and interfaces to appropriate physical client nodes.

Identify data that will reside on each node. Consider the issues associated with distributed data. Check that all processes, which require access to data, map against data that exists on that node, or that the appropriate interface exists to access the required data on a server.

Identify all servers that the client nodes must access and note the data volumes and frequency that the data access will require.

Re-examine the communications network design to ensure that there are no adverse effects caused by data and process placement design decisions.

CraigBorysowich (CTT)