move potentially long running and blocking requests into workers or blocking sections

[chris-giblin]

During processing of encryption decryption requests, it is possible that a call to a network-based service such as for policy or key retrieval is possible. Despite efforts to cache, these blocking calls can occur and will result in errors due to the vert.x main loop constraints. Therefore the current initial code needs to be revised to take advantage of the patterns available in vert.x to support long-running and blocking code sections.

When properly programmed vert.x errors will be avoided. However, Kafka protocol time outs may be unavoidable.

First discuss in PR #31 (review)