LIFO Cache as Server Architecture
Last in First out cache such that older documents not accessed frequently have their index topic shrunk: content is sent to an archive, only to be retrieved later when needed.
As servers scale upwards, it might be useful to find a way to schedule reduction in hard disk footprint by excising some textual fields from the index and moving them to another place, itself perhaps, as multiply redundant as the Solr indexes, just not tied up with the particular overhead of a Lucene index. 

Trivial thought: replace the field's value with a value which points to the archive of that object; perhaps set a isArchived boolean.

As in any LIFO, if someone fetches the value, then it's back, and some other value is archived in its place.
Immediately related elementsHow this works
-
SolrKnowledgeFederationServer »SolrKnowledgeFederationServer
Concepts and Terminology »Concepts and Terminology
LIFO Cache as Server Architecture
+Commentaires (0)
+Citations (0)
+About