MIAMI, 21 MARCH 2011 - Twitter has wrapped up a major overhaul of its IT infrastructure designed to dramatically enhance the microblogging service's uptime, performance and availability.
After a series of lengthy and disruptive outages last summer, Twitter embarked on a revamping of its back-end systems in September, a project that has now been finished, the company said on Monday.
The overhaul amounted to Twitter's "most significant engineering challenge" and is expected to have a significantly positive impact, ensuring that the company can "stay abreast of our capacity needs; give users and developers greater reliability; and, allow for new product offerings," the blog posting reads.
If the infrastructure upgrade meets its expectations, it will be a major improvement for Twitter, whose system infrastructure has routinely struggled to keep up with the site's meteoric user growth.
In addition to improving its service to end users and to external developers, Twitter is also helping its efforts to boost ad revenue. Twitter, which is celebrating its fifth birthday this week, launched its Promoted Tweets advertising program almost a year ago with the expectation that it will allow the company to sustain its business with revenue that is proportional to its massive popularity.
Last week, the company said that while it took more than three years for users to send the first 1 billion "tweets," that now happens every week. Moreover, the average number of Twitter messages sent per day has increased from 50 million in March 2010 to 140 million this month. An average of 460,000 Twitter accounts are created every day, while the number of Twitter mobile users has spiked 182 percent year-on-year.
Among the main improvements Twitter implemented are the ability to replicate user posts, or "tweets," to multiple data centers -- about 20TBs worth of data -- as well as the creation of tools and processes to better monitor its systems, including servers and networking equipment.
"With hardware at a second data center in place, we moved some of our non-runtime systems there -- giving us headroom to stay ahead of tweet growth. This second data center also served as a staging laboratory for our replication and migration strategies. Simultaneously, we prepped a third larger data center as our final nesting ground," Monday's post reads.
Adding to the complexity was that the makeover had to remain transparent to end users and not impact the service, even while usage climbed, and that a number of unrelated new products and features were launched during this time.
"This move gives us the capacity to deliver Tweets with greater reliability and speed, and creates more runway to focus on the most interesting operations and engineering problems," the blog post reads.
Sign up for Computerworld eNewsletters.