Lan Huang^1,, Juan Zhou¹

Microblog services, such as Twitter and Weibo in China, has become a new yet powerful information dissemination channel. More than 500 million tweets are sent every day. The extraordinary large number of messages brings new challenges to conventional search paradigms: a message might be relevant to the query in many aspects, for example the content, time and location of a message. Furthermore, there might be a large number of such relevant messages. In order to address these challenges, we designed a multi-facet distributed microblog search system using off-the-shelf open source frameworks including SolrCloud, Hadoop and Zookeeper. The system was then populated with real world messages collected from the most popular microblog website in China: Sina Weibo. We compared the performances of the standalone and the distributed version of the system. Empirical experimental results showed both effectiveness and efficiency of the proposed system in retrieving large scale microblog messages.

solr, SolrCloud, multi-facet retrieval, information retrieval, microblog