The main philosophy behind the architecture of a NetSolve system is the following. Each NetSolve process (instance of the agent or computational resource) is an independent entity. The system therefore can be modified without endangering its integrity, since any NetSolve process can be deleted/created at any time. For instance, it is possible to have a NetSolve system with no agent. Such a system is just not accessible by any user. An instance of the agent can be restarted later on, to make the system accessible again.
Managing such a system can rapidly become difficult, and a tool is needed to have a centralized view of the system. To make this tool as convenient as possible, we developed two CGI scripts accessible from a World Wide Web browser. These scripts take as input the location of an agent instance (the name of the host where it is running) in order to identify the NetSolve system to inspect. The first script outputs the list of agent instances or computational resources available in the system (a list of host names and IP-addresses). The second script outputs the list of problems solvable within the system.