Write the "Storage Engine" RFD

[alilleybrinker]

One of the challenges with Hipcheck is that the scale of data needing to be analyzed can quickly outgrow the capacity of memory, and ideally we want to manage this ourselves to begin backing this down to disk. There's extensive literature on how to handle this kind of caching, and Hipcheck does present one benefit over other theoretical systems, in that we know up front what data is expected to be collected by the workload being run, and can try to make intelligent decisions about how that data will be stored.

Related to the API design, we'll need to provide some mechanism for data to be sent or shared between plugins and the central Hipcheck engine, and we'll have to determine the level of control plugins will be given over that storage and transmission.

The storage engine RFD should reflect how data is stored, what controls are available to Hipcheck users and to plugins to impact that storage, and ideally how storage can be introspected and debugged.