When low cost carrier AirAsia first launched in 2001, business started with two aircraft and some 200,000 passengers yearly. The airline has since expanded to an operation comprising 230 aircraft and 90 million passengers annually.
An organizational restructure accompanied this growth – a new holding company called AirAsia Global was formed that separated the airline and non-airline entities. The latter includes AirAsia’s physical asset business, ground services, shared services and food and beverage services. With the restructure came the need to harness the power of data for the benefit of all businesses under the AirAsia umbrella.
“Our aircraft used to be a key asset for data, and there’s a lot of numbers from that source alone,” said Aireen Omar, deputy group CEO of digital transformation at AirAsia during NEXT 2018 in San Francisco. “But utilizing data from all our businesses is now more crucial than ever, as there are many opportunities to build around it and better serve passengers.”
AirAsia’s has three broad targets for its data strategy. The company is firstly looking to increase revenue by using data to launch more personalized products. Secondly, data is expected to help increase internal productivity and efficiency. Lastly, AirAsia is aiming to use insights gleaned from data to improve the passenger journey from the moment a ticket is booked.
A large part of achieving the above goals involves the cloud. The company is working on first centralizing its data and putting it into the cloud. From there, data sets will be cleaned up before ventures can be made into machine learning.
Going serverless
AirAsia’s journey in the cloud began seven years ago. The airline adopts a multi-cloud strategy, utilizing a number of public cloud providers such as Amazon Web Services, Google Cloud Platform and Microsoft Azure for various purposes. The company uses Alibaba Cloud in China to more efficiently serve the China market since 20% of ticket sales and 50% of its web traffic originates from the country.
While the company has the option to add servers as it grows, it has instead chosen to adopt the hybrid, serverless route to better support its data strategy.
“All our APIs are exposed to data services, and all those APIs run serverless. Our data collection services are also serverless,” said Nikunj Shanti, chief digital officer at AirAsia.
AirAsia now uses Google App Engine to develop and host web applications. The airline runs Google BigQuery to extract, transform and load customer data from its systems, websites and mobile applications for analysis, while Google Data Studio BETA assists with reporting and dashboards. AirAsia developers utilize Google Cloud Pub/Sub to integrate systems hosted on Google Cloud Platform or externally, while Apache Airflow is used to create, schedule and monitor workflows.
The airline has enjoyed some 5 – 10% of operational savings on the data side, and an 8 – 10% increase in conversion rate since the deployments.
Away from Azure
According to Declan Hogan, Group CIO at AirAsia, the company is aiming to have no compute on-premise by the end of next year and has not bought a server in a year and half. “Cloud computing has evolved. It used to be about VMs but we’re looking to skip containers and go straight to serverless to avoid the complexity associated with managing containers,” said Hogan.
Hogan foresees a fair volume of application rewriting on the horizon as the company moves forward with its serverless approach. While AirAsia’s migration to the cloud over the years has involved many ‘lift and shift’ instances, especially for legacy applications, re-architecting applications is inevitable.
“You can’t Docker and containerize some applications on Microsoft, and a rewrite would mean rewriting into .NET which would be far too much trouble, so we might as well re-architect while trying to keep it as simple as possible,” said Hogan, adding the future will likely result in a move away from Microsoft Azure.
“Many of our applications were written in .NET so there was the appetite to go with Azure, but that’s just legacy,” said Hogan. “Having data driven initiatives means we need a re-design.”
The redesign has so far resulted in the IT team decomposing the AirAsia’s booking engine, with a new, serverless booking engine set to launch in Q4 this year. The airline previously had 4 to 5 booking platforms, but has since reduced this to 2. The IT team is also working on rewriting its mobile app in order to take advantage of new, centralized services.
“We’re learning a lot in this development process,” said Hogan. “It’s not merely an IT project; the dispersion of info is critical in making things successful. If more people understand projects, it makes things easier for the whole business.”