Run Client-Server Mode

AcuFieldView allows multi-process operation in addition to the standard (single process) method of using AcuFieldView.

Multi-process operation permits AcuFieldView to split functions that are data and compute-intensive from those that use graphics and provide control (interactive or script-based). The data- and compute-intensive functions are handled by a program called the server and the interactive/control operations are handled by the client. Client-server operations use a separate process for the client and each of the servers. This method of operating allows the optimum use of computing resources that are distributed throughout your organization or the world.

AcuFieldView's client-server architecture is unique in that there is a choice as to whether or not to operate in client-server mode or as a single process. In single process operation, the data is read directly into AcuFieldView. This mode of operation is called direct. Direct mode is selected by default.

When client-server operation is desired, you simply choose a server. Each dataset that is read requires a server process. In most cases, AcuFieldView will automatically start the server process on the specified machine and read-in the data. This is called automatic server start. Other than choosing a server, there are virtually no differences in using AcuFieldView in direct or client-server mode. When a dataset is replaced or AcuFieldView is exited, the server is automatically shut down.

Table 1.
Configuration	Operation	Benefits
Any client-server operation	Remote server on local (client) machine or remote machine.	Gain a separate address space for client and server, so that larger datasets can be read. Separate address space for each appended dataset so that many large datasets can be read.
Client and server (s) on same machine	Select the default remote server (which is actually the client machine). The server process is started automatically during data read.	See above
Client and server on separate machines, access to a remote and a local dataset in one session.	Local datasets are read from the client machine (or any file system visible to the client machine). Remote dataset is appended by first selecting the remote server machine and then using the remote file browser to locate the file to be read.	Remote server access does not require that the file system of the server machine be visible to the client. Data files do not need to be copied from a remote server to the local client. Large datasets consume memory on the server machine, not the client (desktop). Side by side comparison of data that is stored at different physical locations.
Two or more remote servers on server machine	Remote datasets are read in either replace or append mode by first selecting the remote server machine and then using the remote file browser to locate the file to be read.	Gain a separate address space for client and server, so that larger datasets can be read. Separate address space for each appended dataset so that many large datasets can be read. Remote server access does not require that the file system of the server machine be visible to the client. Data files do not need to be copied from a remote server to the local client. Large datasets consume memory on the server machine, not the client (desktop). Side by side comparison of data that is stored at different physical locations.
Two server machines	Remote datasets are read in either replace or append mode by first selecting the remote server machine and then using the remote file browser to locate the file to be read.	Gain a separate address space for client and server, so that larger datasets can be read. Separate address space for each appended dataset so that many large datasets can be read. Remote server access does not require that the file system of the server machine be visible to the client. Data files do not need to be copied from a remote server to the local client. Large datasets consume memory on the server machine, not the client (desktop). Side by side comparison of data that is stored at different physical locations. Each dataset gets its own CPU and memory, thereby increasing performance.