When is it appropriate to burst to cloud from an on-prem HPC cluster? Are there certain metrics and ranges for those metrics that might serve as triggers? What factors go into a decision to ‘burst to cloud’?
A good starting point may be the paper ‘from Laptop to Lambda: Outsourcing everyday jobs to thousands of transient functional containers’ by Sadjad Fouladi et’ al
you may also want to checkout fuzzball
@ Saifi, that is a good one. I personally don’t like clicking on any attachments, so I post the link for people to make their own judgement. In addition to the Pdf. The link below (credit Sadjad Fouladi),
has the lightning talk and the presentation slides. You can also download the pdf (article).
https://www.usenix.org/conference/atc19/presentation/fouladi