According to Twitter Engineering blog, Twitter has enhanced the search speed using Blender, a Java server. This change has improved the results by reducing the search latencies by approximately three times from 800ms to 250ms and also cut-down the CPU load on Twitter’s front-end servers.
Twitter illustrates the advantages of the new search architecture with the recent spike in search queries after the Tōhoku earthquake and tsunami hit Japan on March 11, which significantly increased search latencies. After the Blender is deployed, the average search latency dropped to around 225ms and is stable since then.
The bottom line is that for Twitter it means fewer servers and fewer costs whereas for end users, twitter’s search engine will work faster which is currently serving more than 1 billion queries a day.
