Accelerated Data Centre Exit Unlocks Agility and Scalability

Home / Resources / Case Study / Accelerated Data Centre Exit Unlocks Agility and Scalability
Case study - Accelerated Data Centre Exit Unlocks Agility and Scalability

About the Client

Skyscanner is a leading global travel search engine. More than 80 million people use it every month and its award-winning app has been downloaded over 80 million times.

We were enlisted to help Skyscanner complete its migration onto Amazon Web Services (AWS) as data centre closure loomed. Working closely with Skyscanner’s product teams, we devised and implemented a migration strategy which completed on schedule, avoiding a £5 million cost that could otherwise have been incurred in data centre contract and hardware renewals.

Challenge: Cloud-native Migration Taking Too Long

The online travel industry is highly competitive, and Skyscanner has a demanding global customer base. Any downtime caused by IT failure could result in reputational damage and lost revenue.

To meet escalating and evolving needs, Skyscanner needed to scale and bring new products to market quickly without compromising stability. However, the operational constraints of physical data centres were putting a stranglehold on development and scalability. What’s more, high maintenance costs were absorbing funds that would have been better spent on innovation.

While an all-in move to the cloud was necessary, Skyscanner understood from the outset that it would not be an easy undertaking. Using a squads and tribes model, Skyscanner’s services are built and managed by highly autonomous product teams. Within engineering guidelines, teams are empowered to choose their own ways of working, the technology they use and the direction their products take. They were also free to select their migration approach.

Initially, all teams chose to focus on rebuilding applications so they could become truly cloud-native on the AWS platform. This approach meant they gained considerable value from early migrations. However, software rebuilds are extremely slow, and after 18 months, it became clear that the full migration would not complete on schedule if it continued at this pace. The business would be left with dual running costs for an inordinate amount of time, and there was no incentive for teams to migrate ‘long tail’ products not being actively developed.

“[This collaboration] has been instrumental in getting our cloud migration and datacentre exit plans running on schedule. I have been extremely impressed by the quality of [the team’s] work and the experience they bring. They’ve been a crucial enabler to us hitting our strategic deadlines,”

Ashley Sole
Senior Engineering Manager at Skyscanner.

Solution: Pragmatic Approach Speeds up the Process

A dedicated team of our consultants and engineers devised and implemented a strategy to accelerate Skyscanner’s migration to AWS. They actively collaborated with Skyscanner’s product teams to ensure objectives and priorities were properly understood.

Most services migrated by the team followed one of two migration patterns: they were either repackaged into Docker containers and deployed using Skyscanner’s proprietary toolchain; or they followed the lift and shift route.

Those that were moved into Docker containers were deployed into Amazon ECS using Skyscanner’s proprietary Slingshot tooling. This facilitated ease of integration with the standard platform service, including centralised secrets management and Kafka log aggregation. Some rearchitecting was sometimes involved, for instance evolving stateful components to introduce statelessness.

Services migrated via lift and shift were principally Windows systems running .NET and SQL. If time had permitted, many of these would have been rewritten or deprecated. But in a deadline-driven context, this work wasn’t a high enough priority, so they had to be moved as they were. We used AWS’ Server Migration Service (SMS) for these migrations, extending open-source tools to automate the creation of CloudFormation templates as part of the process. That gave us infrastructure-as-code and improved cloud-native operability, despite following the simple lift and shift route.

Thanks to our support, Skyscanner completed its server migrations without having to renegotiate or extend data centre hosting contracts and is now 100% hosted out of AWS.

“[The team] provided Skyscanner with a flexible resource pool to help finalise our transition from datacentres to AWS. They provided skilled engineers with the AWS knowledge required to complete a number of our migrations.”

Peter Sturrock
Senior Director Engineering, Skyscanner

Outcome: Migration Completes Ahead of Data Centre Closures

Cost avoidance

If the migration from physical data centres had overrun, a substantial hardware refresh would have been unavoidable. The hardware in many of the data centres was approaching end-of-life and incidents related to this were becoming more frequent. Completing the move on schedule ensured dual running costs were avoided, saving an estimated £5 million in contract renewals and hardware costs.

Improved focus on innovation

For Skyscanner, one of the most noteworthy benefits of deploying in the cloud is the release of previous restrictions on scalability. Product teams are newly energised and empowered to focus on products and services that improve competitive differentiation and customer satisfaction without having to focus on the ever-increasing, scaling demands of a rapid-growth product. Working with modern cloud services facilitates the development of innovative new solutions without the heavy lifting. The service-based model supports ‘fail-safe’ working practices, enabling the product teams to experiment without incurring large or protracted running costs. Importantly, speed-to-market is improved as infrastructure services can be provisioned in minutes, rather than days.

Speed of migration

Partnering with us helped Skyscanner achieve its goal of migrating from its five global data centres to AWS by the end of 2018. The attitude and aptitude of our Agile and cloud-native engineers meant we dovetailed effectively with Skyscanner’s delivery teams to accelerate change.

“[The team] adapted well to the fast-paced environment of Skyscanner. Their agile mindset brought clarity and objectivity to a highly complex cloud migration. Their skills and experience allowed us to focus our efforts and make quick wins to achieve a strategic deadline. They ultimately helped us to achieve timely completion of the migration.”

Ashley Sole
Senior Engineering Manager, Skyscanner

This case study is based on work completed by DevOpsGroup before the team joined forces with Sourced Group, an Amdocs company.