Not a magic bullet
- Can multiple the number "machines"
- At scale not as efficient as running single OS
instance on bare metal...
thrash your level 2 cache, etc
- You still have to admin the machines.. even up with
more to admin, not less
- dev -> QA -> production can be the same image
- great container... lets you pre-stage and do large
scale atomic switches and rollbacks
- useful for HA with VM motion?
- provides a good layer of abstraction... hardware
changes, machine "image" stays the same
- great if you have servers you want to version
separately, but they won't consume resources on a single machine
- motion for zen