I need to setup a testing system for my company. Currently we are doing the following:
- Build all baselines (latest, stable, customised1, customised2), just the binaries
- Test those binaries on two times two systems (Windows-7 and Windows-x)
The testdata are equal on every system (which is not desired).
Therefore I'm looking to solve the following situation:
- Build all baselines, based on GIT branches.
- Create installers for all baselines (configuration might depend on baseline).
- Install and test each baseline/platform individually (testdata might depend on baseline).
This however comes down to the following table:
| Latest | Stable | Customised1 | Customised2
--------+--------+--------+-------------+-------------
W7 | x | x | x | x
Vista | x | x | x | x
WServer | x | x | x | x
W8 | x | x | x | x
W10 | x | x | x | x
In order to perform those tests, until now we've always been working with virtual machines over an ESX server.
As you can derive from the table, we would need up to 20 virtual machines, and I'm wondering is there any kind of technology which makes it possible to reduce this number? (Is there any kind of technology which can emulate different Windows-x PCs and execute the tests for the different baselines simultaneously?)