Malaysiakini Delivers Fast, Timely News to 1M+ Concurrent Viewers During Peak Periods at Half the IT Cost

With an internet penetration rate of 88 percent and growing, online media is gradually overtaking TV and print news as the main source of news for people in Malaysia. Malaysiakini has been a pioneer of online independent news in the country since 1999. As a fully self-funded media outlet, Malaysiakini’s mission centers on providing fast, accurate, and independent news, alongside well-informed and diverse views. Its platform serves more than 10 million monthly readers with content published in Bahasa Malaysia, English, Chinese, and Tamil.

With technology as its backbone, Malaysiakini was an early adopter of cloud computing. The company has used various cloud platforms in the past decade before migrating to Amazon Web Services (AWS) in 2018 for increased cost control, scalability, and enhanced transparency.

Service limits from other cloud providers that led to downtime during high traffic events prompted the move. On an average day, the Malaysiakini site serves around 12,000 concurrent viewers. However, during peak periods such as national elections, that figure skyrockets to over 1 million concurrent site visits. In addition, its engineers became frustrated with a lack of visibility and granular controls for out-of-the-box services on other providers’ platforms.

“We never knew how our services were running in the backend,” relates Aizu Ikmal Ahmad, chief technology officer of Malaysiakini. “On AWS, it’s easy to understand the underlying components for each service, and we can simply run the ones we want according to fluctuating business requirements.”

Malaysiakini promotes an innovative, cost-saving culture from the bottom up. The ability to constantly experiment and adjust settings for cloud services is essential from an IT perspective. “We operate like a non-profit and are always experimenting with our architecture to strike a balance between ease of engineering and service costs,” Aizu elaborates.  

Through constant “tinkering” and granular control of its cloud resources, Malaysiakini reduced its infrastructure costs by 50 percent on AWS. To minimize compute costs during peak news periods, Malaysiakini sets up static websites using Amazon Simple Storage Service (Amazon S3). Traffic hitting the primary dynamic news site, malaysiakini.com, redirects to a custom static page at live.malaysiakini.com.

“Amazon S3 offers high scalability and availability, as well as cost-efficient configurations. Our engineering team wants to handle high traffic volumes cost-efficiently and our editorial team wants a custom webpage for big events. It’s a win-win for both teams,” says Eshwaren Manoharen, senior engineer at Malaysiakini.

Malaysiakini runs its main site on Amazon Elastic Compute Cloud (Amazon EC2), but has offloaded non-essential services requiring high bandwidth, such as its Really Simple Syndication (RSS) feed, to Amazon Lightsail virtual private server (VPS) instances. The Malaysiakini website consumes 1 TB of data transfers each day—about 10 times what typical news sites consume—so offloading traffic where possible to Amazon Lightsail and Amazon S3 has proven a key strategy in cost control.

“With Amazon Lightsail, we get a fixed amount of bandwidth every month, which is incredibly valuable for a publishing business,” Eshwaren says. Applications running on Amazon Lightsail have maintained near 100 percent uptime over the past year, improving readers’ retention and engagement.  

A robust DevOps practice is also behind the success of the dynamic Malaysiakini news site. Previously, Malaysiakini was using third-party services to build and deploy development pipelines. However, this approach was slow and caused several hours of downtime during software updates. Small changes to the site—such as switching the color of a button on a homepage—took 20–30 minutes.

On AWS, Malaysiakini no longer experiences downtime and updates take an average of 7 minutes, from writing to deploying code in production. This has significantly decreased the response time for software updates, enhancements, and bug fixes on the Malaysiakini news site.

Malaysiakini uses AWS Elastic Beanstalk to automate infrastructure deployment from its DevOps pipeline, and AWS CodePipeline as a fully managed continuous delivery service. AWS Elastic Beanstalk keeps Malaysiakini’s application up to date with automated, ongoing patching and updates, so developers can focus on writing code. “The whole setup is really easy, and deployment is much faster and more stable on AWS,” Eshwaren adds.

On the business side, Malaysiakini appreciates the ease and transparency of billing on AWS. “Accounting is simpler because our billing isn’t fragmented across cloud and software providers,” explains Aizu. When looking to build a new feature, engineers typically opt for AWS-native services, which simplify billing and system integration.

Recently, the company started sending a newsletter to subscribers, for which it’s using Amazon Simple Email Service (Amazon SES). “We tried other email services and found Amazon SES better in every way, with faster message delivery and lower pricing. It reinforced why we decided to migrate to AWS in the first place,” Aizu says. “Many third-party solutions use AWS services in their core tech stack. The advantage of running our infrastructure natively on AWS is that we can replicate or reverse engineer most out-of-the-box solutions and run them at a lower price.”

Malaysiakini aims to continue expanding its subscriber and advertising base by reporting on the latest current events in Malaysia and across the globe. The company also takes on external media projects as needed to ensure a continuous supply of independent funding. A recent project, for example, was the creation of the EPIC Tracker search hub—a front end document repository for public health, economic, social, and community measures taken to address COVID-19 across all countries—for United Nations University’s International Institute for Global Health.

“When we mention to new clients that we’re running on AWS, it gives us a lot of credibility and is a strong selling point,” Aizu says. “It’s also reassuring that everything works out of the box on AWS. Nothing requires custom configuration, which saves us time and helps us follow consistent development and deployment methods for every client and project.”

To learn more, visit aws.amazon.com/media/