| Customize Help

Distributed MIL overview


Distributed MIL (DMIL) allows a MIL application running on one computer to access or use the resources of other computers, for example, to distribute processing, to grab remotely and display locally, or to monitor their buffers and other objects.

There are two Distributed MIL configurations:

  • Controlling configuration. A Distributed MIL setup comprised of a single MIL controlling application that uses DMIL to allocate MIL systems on remote computers to share in processing and other MIL operations; the DMIL remote systems can collaborate with each other and with local systems. You would typically use this to distribute the processing of a MIL application among several computers.

  • Monitoring configuration. A Distributed MIL setup comprised of a single MIL monitoring application that uses DMIL to monitor published MIL objects (for example, buffers) of one or more independent MIL publishing applications; the MIL publishing applications use DMIL to publish designated MIL objects. You would typically use this to create a MIL application on a local computer that monitors MIL applications running on remote computers.

In both configurations, a group of computers interacting using Distributed MIL is called a cluster. A cluster can have up to 128 computers. Regardless of the configuration or whether running locally or running remotely, all MIL functions in a Distributed MIL cluster generally behave as they do when running on a single computer without Distributed MIL.

All information in this chapter applies to both configurations, unless otherwise specified.

Setup and installation

To develop or run a Distributed MIL application, MIL/MIL-Lite must be installed on all computers in the cluster. All computers should have the same version of MIL installed, including Service Packs and Distributed MIL Updates, with appropriate licenses; see the Licensing considerations subsection of the Preparing computers for Distributed MIL section later in this chapter.

The operating systems of computers in a cluster can be a mix of any MIL supported operating systems, including 32-bit and 64-bit versions. If exchanging buffers across the network, it is recommended that the computers have at least a Gigabit Ethernet interface (1000BaseT). The connection between computers is generally made using TCP/IP with a client-server architecture.

Client-server architecture

Both Distributed MIL configurations use a client-server architecture. In a client-server architecture, a client makes requests to a server, which accepts the requests and returns a service. When you install Distributed MIL, the Distributed MIL server is installed.

In the controlling configuration, the controlling application is the client, and it allocates the DMIL remote systems on computers running a Distributed MIL server. You must explicitly ensure that the Distributed MIL server is running on the remote computers; although by default, the Distributed MIL server is automatically running. For more information on managing the Distributed MIL server in the controlling configuration, see the Setting up the Distributed MIL server on remote computers subsection of the Preparing computers for Distributed MIL section later in this chapter.

In the monitoring configuration, the monitoring application is the client and the publishing applications have Distributed MIL servers loaded in them. In this configuration, you do not directly manage the Distributed MIL server; it is transparently managed by the MIL functions that establish the connection between the monitoring and publishing applications. For more information on which functions manage the connections in a monitoring configuration, see the Monitoring configuration section later in this chapter.

Typically, you use Distributed MIL to access or use the resources of other computers. However, there are occasions when it is useful to run a Distributed MIL configuration only using your local computer. For example, when developing or debugging a Distributed MIL cluster, it might be useful to test your cluster completely locally before deploying it across several computers, where you have to worry about possible connection issues. For more information, see the Distributed MIL on your local computer section later in this chapter.