Hatstall scalability

The current hatstall design does not scale. When there are more than 10,000 identities in the SH database, the performance is not good enough. And with 100,000 identities is unusable. In this ticket a new design will be proposed and a plan for its implementation defined.

* Instead of loading all identities in SortingHat, only a limited number of identities will be loaded. Then, with pagination, the rest of the identities will be loaded.
* The list of identities to be loaded will be defined using some kind of searching/filtering, so in general, the pagination won't be done in all identities in SortingHat database but only in a subset.

This is just a proposal to be discussed and it will evolve to final design that will be implemented.

No files, tests, or entry points are named. Start by reviewing how Hatstall currently loads identities from the SortingHat database, then define the proposed search/filtering and pagination design. Done means an agreed design and an implementation plan for scaling beyond 10,000 identities.

Written by the indexing model from the issue text.