Full-text Search Engine and Document Delivery System in an IC Client Site
This site delivers 10 million+ pageviews a month. Our customer who owns this site started operations in 2007. The website primarily serves documents. The website primarily serves more than 200 thousand documents stored in SQL database. The site has a powerful full-text search feature built on top of the FAST search engine. FAST, indexes the documents in the database and stores them in multiple index shards. Each shard or fragment is of a manageable size. Searches are parallelized and handled by search servers with each server handling a portion of the index. The results are merged, ranked and served by the web servers. System architecture is akin to and inspired by Google.