You can distribute your data in different databases and combine them with views. The hard part of this will be to update the views when adding / removing data.
You need to make this "section" on a logical key (for example, in a calendar date), where each database has data in a certain range. If you cluster in this field, the query analyzer will be able to determine which database to output data without problems.
At my workplace, we use this method for a very large (multi-billion dollar) data set, with which we receive monthly additions, and it works great.
source share