Chris, there seem to be two camps firming up around hybrid. Some say that cloudbursting, for example, is just too complex and often infeasible given WAN limitations, and that if you think an app will be bursty you should just run it in the public cloud. Databases that must stay behind the firewall are another problem.
Others think it is feasible to shuttle workloads between a private cloud in a colo or enterprise data center and AWS/Google etc. -- it just takes some effort, will, and automation chops. They think this back and forth will get easier and companies can start positioning workloads fluidly based on cost and other fluctuations.
What do you think?