> The faster startup time of VMs is cool and I appreciate the work put into the paper... I'm just saying that it doesn't seem to matter in the bigger picture.

> The faster startup time of VMs is cool and I appreciate the work put into the paper... I'm just saying that it doesn't seem to matter in the bigger picture. I imagine this kind of analysis is aimed at use cases like, say, AWS Lambda, where you're launching containers/VMs on something like a per-request basis.