I have the following need:
- Simulate 1K (or more) web clients (i.e., http requests) to an ip address.
- Each client has to use a different mac as source.
- Solution has to be somewhat realistic, i.e., i am hoping to get (under fundamental limits) maximum simultaneous requests from this clients.
- All this in the same machine
- Using linux.
My Question: How can I do this maximising the number of users and requests per second?
I am quite certain (99% :)) of what I am doing. I really need these requests to be send with a different mac address as source since it is a special use case based on OpenFlow switches. But it is irrelevant to expand this scenario into the question.
P.S. I am currently working in python using the Eventlet Networking Library. Python/Eventlet is not a requirement, but I include it here since i have a certain urgency and a fast implementation/modification of my code would be beneficial. I was thinking that maybe i could associate virtual interfaces with different mac addresses but i am lacking a way to specify the request interface.