SRF delivers fast and reliable realtime news updates to millions users

SRF operates a live ticker platform for its millions of users, who use the site to find out the latest news and current affairs. For select stories the live feed is used to provide news updates, as they happen, which users can subscribe to and get alerted when something new is added. SRF prides themselves on fast, up to date content to keep its audiences informed as stories unfold and progress. But it is hard to predict how many users will be logging into the site at any one time, especially since stories spark varying levels of interest - and as a result varying levels of traffic. No matter what, the infrastructure has to be ready to handle traffic influxes for times when posts are (sometimes unexpectedly) popular - or even go viral.

SRF were using long polling, built within their own infrastructure, to manage their news feed updates. The frontend app requested updates from the server approximately every few seconds. In low traffic scenarios this wasn’t such a problem - but for popular stories, where lots of users were reading the news feed, the servers were hitting capacity. With the risk of diluting the user experience, SRF knew they needed to look for a new solution.

SRF wanted to move from long polling to a realtime WebSocket solution. They initially developed a proof of concept within their own infrastructure, but it became clear that they would have to invest in more dedicated resources to manage and maintain it themselves. 

Instead of taking the work on, which would have delayed the deployment of the improvements, as well as introducing overhead, they looked to Ably to provide a SaaS infrastructure solution.

Hoping to get the enhancements to market quickly, SRF were pleased to find the development with Ably very straightforward. Urban Etter, Technical lead at SRF said, 

“Ably was easy to work with – documentation was really helpful and direct contact with your developers made reaction times really quick.”

SRF uses Ably to manage the live feed with content being pushed to the user in less than one second. For big, fast moving stories, this enables an editor to type an update into the CMS, which is first saved to the database before being published via Ably - where the news update becomes immediately available to view. 

When an update has been published, a message appears at the top of the SRF newsfeed, notifying the readers. The update button pops up in realtime, and once clicked, dynamically updates the page rather than reloading it. This enhances the user experience as simply injecting the live news update as it is uploaded can make the pages jump while reading. 

Using Ably has proved invaluable when big stories have hit the channel. Historically, on an average day, traffic and messages range between hundreds of thousands, but recently there was a big federal election in Switzerland that generated huge interest. 

At its peak, the SRF live ticker had 62,000 concurrent users connected, all receiving the updates at the same time. Over the course of the day traffic reached 20 million. 

Seeing such huge spikes in traffic would ordinarily cause developers some concern but this is where Ably’s elastic scaling infrastructure really came into its own. It was able to easily handle the rapidly changing numbers of devices connecting and absorb the very large, unexpected spikes, with autoscaling systems that actively managed the capacity. 

The first indication SRF had of the load was an email from Ably letting them know that there was increased activity on the platform. As explained by Urban, “We didn’t notice anything! The servers were really calm, which is a very good thing.”