Apache Blur
Blur combines the speed of document-oriented databases with the ability to build rich data models to deliver lightening-fast answers to complex queries
http://incubator.apache.org/blur/ There are a number of ways in which one might consider Blur in an Open DeepQA project. Since it is based on sharded Lucene, it performs many of the important functions of Solr.
In one case, Solr might be replace by Blur.
In another case, Solr might host topics as individual SolrDocuments, but persist bulk information resources related to those topics, e.g. long descriptions, in Blur, allowing the Solr index to remain as compact as possible.