Blog

Ellen DeGeneres Selfie Twitter Crash Should Serve as Wakeup Call for All Other Apps

March 5, 2014 by Ixia Blog Team

The world went crazy sharing the selfie of Oscar host Ellen DeGeneres taken during Hollywood's annual Academy Awards ceremony on Sunday, which included stars like Meryl Streep, Jennifer Lawrence, and Bradley Cooper. The unique idea quickly became the most shared photo ever, taking many by surprise – including Twitter, which crashed under the pressure of heavy re-tweets and shares. This is, however, not the first time a popular app stood red-faced. WhatsApp too crashed a few days back, seemingly caused by a flood of subscription requests after the news of its acquisition.

Group pic

It wasn’t all smiles for everyone

Today WhatsApp has around 450 million subscribers, while Twitter enjoys subscription from 645 million plus users. With the popularity of these apps, people are quite regular visitors. In fact, it is claimed that around 70% of the WhatsApp users use the app on a daily basis. So we would think that these apps should be more than capable of handling the scale of such requests, but recent events reflect otherwise.

It seems like these applications are quite capable of handling a certain scale of users simultaneously and performing different types of actions, but what happened differently that could have caused the crashes?

Although we may never know what exactly the point of failure was, in both cases the crash occurred when millions of users were trying to perform a common action. There is a chance that the multimillion execution of single repeated action, like “Subscribe to WhatsApp” or “Share a Picture,” stressed a particular path of the application and brought it towards its breaking point.

The traditional scalability testing done for vanilla HTTP and HTTPS on application servers are inadequate in finding such issues, which calls for application-specific scalability and resiliency validations. This is where Ixia’s BreakingPoint system can help. It supports scalability and resiliency validation of more than 240 applications, including WhatsApp and Twitter. By emulating millions of unique clients executing particular actions against an emulated or real appserver, the tool helps in creation of such unique scenarios. The application scalability testing can be performed in a silo against individual server farms and the cumulative results can be collated to understand the maximum simultaneous actions that an application can sustain. These result in optimizations and better provisioning.

BreakingPoint Emulation

BreakingPoint emulating millions of client actions against server farms

In the future, such use surges won’t be unusual for popular apps and we would do well to change our traditional testing style. Performing application resiliency and scalability testing and taking the right measures ahead of time will ensure continuing service when such incidences happen.

Additional Resources:

Ixia’s BreakingPoint