Feeds

Twitter survives election after Ruby-to-Java move

Peak traffic of 874,560 Tweets per minute without Fail Whale coming up for air

3 Big data security analytics techniques

Micro-blogging site Twitter experienced record traffic as the results of the 2012 US Presidential election were announced on Tuesday night, but the service never faltered despite the increased load – something Twitter engineers credit to the company's move from Ruby to Java for its backend software.

According to a blog post by Mazen Rawashdeh, Twitter's VP of infrastructure operations engineering, Twitter users posted an average of 9,965 messages per second between the hours of 8:11pm and 9:11pm Pacific Time.

During a single second at 8:20pm, Twitter users produced 15,107 new posts, Rawashdeh writes, and during the peak traffic period of the evening they generated 874,560 posts in a single minute.

Such numbers are unusual for Twitter, Rawashdeh says, and they represent an evolution in how customers use the service. While in the past Twitter has experienced brief traffic spikes to mark particular events, such as New Year's Eve or the close of a sporting event, this was the first time such heavy traffic lasted for such a sustained period. It was also the highest volume of traffic during an election since Twitter launched.

In the past, Twitter users might have expected to experience problems when the service is under such heavy load. Early in Twitter's existence, service outages were so common that the company's whimsical "fail whale" error screen became a cultural icon for the Web 2.0 crowd.

But Tuesday night's election-driven traffic went off with nary a hitch, which Rawashdeh attributes to Twitter's ongoing migration from backend software written in Ruby and the Ruby on Rails framework to a new software stack built on the Java Virtual Machine (JVM).

Fail whale

Fail Whale sightings were a lot more common when Twitter ran on Ruby. With Java, not so much

Twitter first began stepping away from Ruby in 2008, when the company's Ruby-based message queuing system "hit a wall," in the words of former Twitter developer Alex Payne.

"There's a lot of things that Ruby is great at," Payne said at the time, "but long running processes? Particularly memory intensive ones? Not so much."

Twitter's solution was to migrate some of its Ruby code to a new server stack running on the JVM. Initially, the company's development team avoided stock Java in favor of Scala, an alternative JVM language that combines aspects of object-oriented and functional programming. Today, Twitter's software is built from a mix of Scala and ordinary Java code.

A few Twitter services still run on Ruby, too, but according to Rawashdeh these are used increasingly less often. Notably, he says Twitter has now reconfigured its system so that traffic from mobile devices never touches any Ruby-based software at all.

Where Twitter does use Ruby, it deploys the code on a custom, highly optimized version of the Ruby runtime designed to manage memory more efficiently when executing long-running processes.

None of this will be welcome news to the army of fanatical Ruby developers who believe the language's syntax, its high developer productivity, and its overall philosophy far outweigh any performance disadvantage it might have compared to other languages.

But for Twitter, results are what matter. "The bottom line: No matter when, where or how people use Twitter, we need to remain accessible 24/7, around the world," Rawashdeh writes. "We're hard at work delivering on that vision."

They had better be. Judging by the torrential outpouring of comment on all things related to the recent election, Twitter can only expect its average traffic load to keep climbing as Barack Obama enters his second term as US President. ®

SANS - Survey on application security programs

More from The Register

next story
Virgin Media so, so SORRY for turning spam fire-hose on its punters
Hundreds of emails flood inboxes thanks to gaffe
A black box for your SUITCASE: Now your lost luggage can phone home – quite literally
Breakfast in London, lunch in NYC, and your clothes in Peru
AT&T dangles gigabit broadband plans over 100 US cities
So soon after a mulled Google Fiber expansion, fancy that
AT&T threatens to pull out of FCC wireless auctions over purchase limits
Company wants ability to buy more spectrum space in auction
EE & Vodafone will let you BONK on the TUBE – with Boris' blessing
Transport for London: You can pay, but don't touch
NBN Co plans fibre-to-the-basement blitz to beat cherry-pickers
Heading off at the pass operation given same priority as blackspot fixing
NBN Co in 'broadband kit we tested worked' STUNNER
Announcement of VDSL trial is not proof of concept for fibre-to-the-node
Google eyes business service in latest Fiber trials
Lucky Kansas City buggers to host yet another pilot program
Huawei exec: 'Word of mouth' will beat Apple and Samsung in Europe
World Mobile Telephone Factory No.3 won't fling the big bucks around just yet
prev story

Whitepapers

Securing web applications made simple and scalable
In this whitepaper learn how automated security testing can provide a simple and scalable way to protect your web applications.
3 Big data security analytics techniques
Applying these Big Data security analytics techniques can help you make your business safer by detecting attacks early, before significant damage is done.
The benefits of software based PBX
Why you should break free from your proprietary PBX and how to leverage your existing server hardware.
Mainstay ROI - Does application security pay?
In this whitepaper learn how you and your enterprise might benefit from better software security.
Combat fraud and increase customer satisfaction
Based on their experience using HP ArcSight Enterprise Security Manager for IT security operations, Finansbank moved to HP ArcSight ESM for fraud management.