I don't know if this forum is a suitable one to my question, I will ask my question and I hope to accept my apologies if you find that it is not suitable....
I have a network with the following topology:
- N hosts (nodes), and One server, and there are links between the server and hosts, (we don’t take in consideration routers or switches) . We can consider that the delay is 0.
- Each host has a set of activities, it can send a request to a server, send a message to subscribe or to unsubscribe, and so on, in total each host has 5 types of actions. The host send a request to a server (to ask it to do something, depend to the type of action) and size of request (number of packets) is depend to the type of request, for example a request to update some information may be larger ( in terms of packets) than a request to unsubscribe….
- A server is in a service of all hosts, so it executes for each type of request its adequate service, and it receives and put each type of request in a particular queue ( there is a queue for each type of service )…
Ps. the politic of scheduling is FIFO
My question is:
- How can I calculate the load of my network at the instance t? ( to test the congestion of my network)
- is the load of network means the amount of traffic ?
- How can calculate the amount of work on a server at the instant t? ( to test the bottleneck)

Thanks for every response and I apology again...