We are building a product that samples various statistics from windows boxes in a domain. The product is similar in architecture to enterprise performance monitoring systems (although we are not doing performance work).
The tricky bit is that we'd like to run our software in an environment with lots of variation in terms of end user hardware, end user installed software, and end user connected peripherials. We already have a test environment that is built from physical PCs and virtualized windows images. However, as a scrappy startup we don't have the ability to build as large a test system as we would like.
Some ideas we have considered:
- Using a lab from a large company. I seem to recall that Microsoft has public testing labs but I can no longer find any documentation. Something like the Apple developer compatibility lab (but for windows).
- Buy a bunch of old machines from ebay. This is a possibility - however the problem is that most sellers wipe the machines clean (thus removing all usage data we are looking for).
- Find a guinea pig willing to alpha our software. Perhaps a school....
Any other ideas / insights?