Having ASCs to compare the performance of VMs is nice. But the impact of the delta remains fairly untouchable. Currently we run a few test series to assign a clear value to the existing performance deltas. But also such a value will not bring clarity.
A clear word is spoken by prices. Assuming you would want to run a bulk compute job in an analytics project, it is necessary to compute as fast as possible. If the amount requires a scale out and more than only a handful of machines, you will be happy to review the MASC.
Assume you have a huge pile of containers, say 10 - your amount of work - to move. And you have two machines to choose from: a small pickup, allowing you to transport one container per run and a large road train, allowing you to complete the job in one run. No doubt you would want to choose the road train if it would not cost a fortune.
And this is what the MASC is about: MASC stands for Meta ASC and expresses a defined amount of work, respectively the time it requires to resolve it given a certain amount of ASCs. We use the EUR per MASC to identify the real costs associated with a VM.
The following diagram outlines the change in price ranking when considering the MASC.