The change has reduced search latencies by approximately three times (from 800ms to 250ms), Twitter claims, and CPU load on Twitter’s front-end servers was cut in half. 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 deployment of Blender, average search latency dropped to slightly more than 225ms and remained stable ever since. A detailed technical explanation of how Blender works is available here. In simple terms, for Twitter this means fewer servers and fewer costs. For the end users, Twitter’s search engine, which serves more than 1 billion queries per day, will work much faster, with the dreaded fail whale making fewer appearances.
Topics to follow
