Each group of servers is composed of either one E3 Server operating isolated, or two E3 Servers in a Hot-Standby configuration. Actually, each E3 Viewer connects to a Domain rather than to a single server.

After connecting, server and client exchange information and check for the existence of registered and updated system objects on the client computer. If such objects are not found, the server uploads object definitions, such as user libraries, according to the need to open each Screen.

From a client's point of view, the download starts on the first Screen (graphical interface), including any internal objects that are saved to a cache directory. This process can take a few seconds, depending on the application, but it brings more benefits next time the user opens the same Screen, because E3 Viewer performs disk and memory cache. If an E3 Viewer is not closed, by the second time a Screen is opened (and it is not reconfigured) it is already in memory, thus reducing pagination time. If a Screen is modified, then E3 Viewer must download it again and restart the process.

After opening a Screen, the server and the client exchange only real-time data and queries to the database using a TCP/IP message system, with high performance when compared to Terminal Services, depending on the application. This is possible because while an E3 Viewer receives and sends only real-time values and data, a Terminal Services client sends keyboard and mouse messages and receives images, generating a higher information traffic. An E3 Viewer connected to an isolated Domain can get information residing only on the active server, according to user access permissions.