Data storage for a WD-schema running process

Refactor the WD-core code for these changes:
1) a data storage should be created for one running workflow process
2) a data itself must not walk through the workflow data-channels, but a special id of the data should
3) the storage is an abstract interface and many implementations of that could exist (in memory, dbi-database). It provides putting, getting and deleting a data using data ids.