Microblogging service Twitter announced that its recently launched search feature has been revamped, TheNextWeb.com reported today. According to Search Engine Land, Twitter receives over one billion search queries a day.
Vator News outlined that users send more than 1,000 tweets per second, amounting to 86.4 million tweets daily. Users embark on 12,000 searches per second, which is roughly 1 billion per day. Search Engine Land also reported that other Twitter benchmarks included April 14, 2009 (when it got 19 billion queries a month) and July 6, 2010 (24 billion queries).
However, "queries" are different from "searches," since a "query" can be done by a machine, such as when a user signs up to receive automatic replies to their tweets. Furthermore, the search is only available for Tweets that are up to seven days old. Twitter told Search Engine Land that it was not planning to provide archive searches since other companies were already doing that and instead would be focused on ameliorating the service in other ways.
The new architecture was speedier and more scalable and only used around 5 percent of the backend resources, which corresponded with the aim to design search "to support at least an order of magnitude more load," Twitter stated. Note: Back-end refers to computer systems that store data and are typically invisible to the user.
"The new system is extremely versatile and extensible, which will allow us to build cool new features faster and better," Twitter stated in a blog post on Wednesday.
The transformation to the back-end had Twitter abandoning the older Summarize architecture and constructing the new search function on open source platform Lucene, TheNextWeb informed. Twitter bought Summarize in 2008 (which was initially an independent company) and TechCrunch wrote that the move hinted that Twitter was planning to branch out into search.
According to ReadWriteWeb, changes to the search engine included:
• significantly improved garbage collection performance
• lock-free data structures and algorithms
• posting lists, that are traversable in reverse order
• efficient early query termination