Provide a highly efficient database for managing complex, distributed task workflows. This is taking the concept of a system bus through message queueing and extending that to also model the concept of one action initiating subsequent actions.

1. Minimise global locking/use lock-free implementations where possible.
2. Ensure transactionability of all actions to guarentee data integrity.
3. Workflows can be manipulated while running.
4. The structure of workflows is dynamic

#Example Workflow ┌──> Task B ─┐ Task A ─┤ ├──> Task D └──> Task C ─┘

##Steps

1. Task A is queued to run
2. Task A is completed which queues Task B and Task C
3. Task B and Task C are run in parallel
4. When the last of Task B and Task C is completed, Task D is queued
5. Task D is completed and completes the workflow