Airboard: Improving time-to-market on AWS, a DNX Startup Case

About Airboard

Airboard is a digital queueing application that removes physical queues to improve the passenger experience at airports and on commercial flights. It currently uses machine learning and its unique patent-pending technology to benefit airports and airlines around the world.

The Business Challenge

As a startup, Airboard had done their homework on the industry, created a great product using agile concepts, and achieved an excellent MVP (Minimum Valuable Product). Airboard was seeking a development team for expedient development (in a two-week timeframe) of a Well-Architected global framework to achieve performance excellence concurrently with high security, reliability, availability, and efficiency for its airline industry customers. A key priority for the digital queuing application is to achieve low latency across multiple, global locations with a highly scalable framework. This requires leveraging the capability of the AWS cloud, anticipating the potential for an exponential increase in the number of simultaneous users as sector adoption grows. Time savings are a significant benefit of the Airboard system, so the accuracy of timing in multiple simultaneous locations remains essential to its success. To achieve these conditions within their desired parameters, Airboard chose to team up with the highly skilled and experienced AWS architects and engineers of DNX to design and build a solution for their first release.

In the initial development phase, the Airboard team were using AWS Lightsail for front-end and back-end PHP applications running on a single EC2 instance, which enabled rapid prototyping in its initial product development phase. However, given the increased sector demand during COVID and as part of a post-COVID recovery solution for the aviation industry, the Airboard team were looking for a way to enable automated deployments that can support global adoption with enough elasticity to allow for spikes in usage during global travel seasons. 

Furthermore, an ambitious customer deadline was imminent for the Airboard team and it was under pressure to prepare the application for its first release. DNX was engaged to not just design and apply a solution for these challenges, but requested by the Airboard team to assist in providing comprehensive documentation and further enhancing its DevOps best-practices on AWS. As a certified DevOps competency AWS partner, DNX pushed hard during knowledge transfer sessions and detailed documentation about our solutions.

At first, going for an event-driven architecture using serverless computing was tempting but required lots of refactoring in the current product at that time, so DNX elaborated a container-based solution on AWS. With critical compliance requirements and strict security concerns, especially in US airports, the due date was close and DNX could modernise the Airboard application while building its AWS foundations.

The Solution

DNX allocated more Cloud Engineers for this project due to its critical deadline, so while a team was building Airboard’s AWS Foundations from the ground up, another one started to modernise the application that was written in PHP with front and back-end separated, both using Laravel Framework and classic LAMP stack (Linux, Apache, MySQL, and PHP). Also, the Continuous Delivery strategy with CI/CD pipelines, essential to fulfilling the customer requirements, started to be designed as the team ran the App discovery phase by the DNX Cloud Architect. 

Our well-known DNX.One Well-Architected Foundation was applied – leveraging our considerable developer experience, and using Terraform to manage our IaC, we could also accomplish high-standard compliance with Airboard’s clients as AWS IAM policies are version controlled and securely managed. Using our IAM topology, the access to AWS accounts are role-based where users assume one or multiple roles across accounts and environments.

Additionally, each policy role has its version tracked using GIT, where any modification or inclusion to a role is approved using Pull Requests. This is a benefit of using IaC, where any change in a policy is tracked and can be compared using git diff.

Application Modernisation

To achieve a cloud-native solution, the PHP application was enhanced with the modernisation process where our engineers review the code and apply 12-factor principles, preparing it for container orchestration on ECS and making sure that performance would not be compromised. 

As a result, we could build the application containers for ECS orchestration, by moving configurations stored in the application to the environments using CI/CD pipelines and ensuring that no state was kept by the application processes. We also automated existing database migrations and deployments that were previously manual processes, providing the team confidence to release new features that can be easily tested in a production-like environment before every deployment.

Continuous Integration and Continuous Delivery

Airboard is a growing business with the foresight to build its foundations on a framework that can scale easily. When DNX were engaged, the team was ready to transition to enhanced pipeline architecture, to support new features and future releases. Prior to engaging us, the Airboard team would connect to the EC2 Instance manually to release new features, as the application was already living in Bitbucket with a pipeline solution. At DNX, we utilised the client’s existing CI/CD tool to provide the best pipeline architecture, focusing on the best approach for the client’s needs. Along with regular feedback, architecture reviews, and Knowledge Transfer sessions, the DNX team designed and delivered a long-term solution to secure Airboard’s scalability in the cloud.

AWS Pipeline

Application pipeline

Some of the AWS Services provisioned:

• AWS ECS on spot instances
• AWS Elastic File System (EFS)
• System Manager
• CloudTrail
• Cloud Watch
• Code Deploy

Customer Benefits

Now Airboard has a future-proofed, scalable solution on AWS with elasticity, global high-availability, CICD, and ongoing automation supporting their application. All infrastructure built in this project uses spot instances that can save up to 70% in costs, maintaining a great Developer Experience. Applying the multi-region strategy created during the AWS Foundation and CI/CD pipelines phase, Airboard can now scale its solution and development team seamlessly around the globe without a significant increase on the current TCO (Total Cost of Ownership), improving passenger experience, supporting the growth of the business, and keeping passengers around the world safe.