I am developing a multi-component network software system designed to work on an arbitrary number of machines. I assume a typical installation of 1 to 4 cars.
I want to take system testing seriously, and I set up a network of virtual machines on a powerful PC that I can use to simulate network connectivity. However, this installation is not enough.
For example, due to virtualization (I use qemu), each node always runs on only one core, so I canβt check the performance issues of code intended for multi-core use. It would be nice if I could sometimes run any of the virtual machines on the whole PC to see what the difference is, especially in terms of performance, and also to check some multithreading problems.
Purchasing more mailboxes and using hard drive partitions instead of images on a virtual disk are options, but are there any more elegant approaches? I am loading a company here and cannot afford to still download equipment, and deal with real physical media for each machine, and not just move image files around, of course, more.
ETA . The application is a middleware system and does not actually have a user interface. Testing is performed using fictitious clients and by feeding data to the system, which is again retrieved at the other end. This is not a website, and it is usually not used over the Internet, but rather inside the local network. User interface and Internet interfaces will be provided at a later stage by some components connected to the system.
source share