Berkeley position paper on the future of cloud computing

I'm still digesting the paper (it's rather long) but working through multi-cloud brokers reminds me of Leslie Lamport's sarcastic description of distributed systems:

  A distributed system is one in which the failure of a computer
  you didn't even know existed can render your own computer
  unusable. [0]

I would like to see a research direction that reduces dependencies in could systems rather than increasing them.

p.s., There have been a number of industrial attempts to commercialize the cloud broker idea. None have succeeded to my knowledge--the market does not seem to be there.

[0] https://lamport.azurewebsites.net/pubs/distributed-system.tx...

Surprised this didn't get more tracking. This in particular seems like a really good startup idea:

> cloud computing will require intermediation, provided by systems we call intercloud brokers, so that individual customers do not have to make choices about which clouds to use for which workloads, but can instead rely on brokers to optimize their desired criteria (e.g., price, performance, and/or execution location).

Perhaps this is the direction startups like 'Mighty' are taking.

Interesting ideas. It would absolutely kill the profit margin of AWS and Azure though. Who's that guy that said "Your margin is my opportunity"?

I’m familiar with the work out of this lab, they fail to realize a couple things.

Vendor lock is real, rarely do resources run in isolation. In organizations, a great amount of work goes into setting up security protocols that are legally defensible. It’s _very_ hard to do this across clouds.

Most cloud abstractions fail because cloud vendors are capitalist entities that get a boost off of lock in. The closest we’ve come is Kubernetes which has managed to get all the clouds on board. This is a rare event, and still has its own limitations.