AWS for Industries
Fast Retailing (Part 2): Building an In-House Engineering Organization
AWS Japan G.K. invited Fast Retailing Co. Ltd. executives to speak on the theme of “Digital Reform Initiatives at Fast Retailing: Transformation of Engineering Organizations Utilizing AWS” on September 2, 2021. This post is part 2 of a two-part series where they share key points of their discussion. You won’t want to miss part 1, where Fast Retailing discusses how they are Transforming the Clothes Shopping Experience.
Fulfilling the Fast Retailing direction to deliver LifeWear to everyone in the world using leading edge technology requires agile and versatile commerce platforms that seamlessly integrate customer and supply chain data. Giving the customer total control of the retail experience meant transitioning our Group Global Unified Platform to in-house development, since this ensures rapid and continuous innovation as well as executes our business strategy.
Fostering In-House Development Success
Our employees perform everything from business requirement formulation to architecture design. We built a system where engineers who understand the contents of the system work together with business units to respond most efficiently to every end user’s request.
In-house development increases delivery speed, flexibility, and scalability while reducing cost. Depending too much on vendors for development often complicates the architecture and implementation, as vendors attempt to implement business requirements accurately. In-house developers can collaborate with the business and create a well-balanced and effective solution.
However, in-house development does lead to some challenges. Hiring developers with the right skill levels is important. You must define roles, processes, and career paths clearly. You will also need to encourage a global mindset across the organization. At Fast Retailing, that meant creating documents in English that developers across global offices could all understand, onboarding new team members according to our values, as well as helping team members in various regions to understand differences in cultures, business customs, and adapt to one another’s unique working styles.
As an organization grows, it’s crucial to create clear team structures for each microservice and assign core people to be leads, members, and partner members. The engineering roles at Fast Retailing can be broadly divided into three types: infrastructure engineer, system architect, and application engineer. Our team culture centers on having a connection with and empathy for the brand; being curious to learn new skills, business issues, and new technologies; and working together as a diverse team to generate results for the business.
We have a hybrid in-house development configuration that enhances the involvement of in-house production according to the importance of a given system. We have prioritized the in-house production of business cores and those that are expected to change or expand.
As a result of years of developing an in-house production system, we can quickly solve new problems and release new services. We work continuously to help our teams accumulate technical and operational knowledge to build more capabilities into the system. Those new capabilities include deploying and operating our Global One solution to all of our brands worldwide, contributing to business needs beyond commerce, learning about large-scale user-based service development, and more.
In-House Development Success Story: Batch Processing Platform and Data Hub Platform
To support Fast Retailing’s direction to position the customer in total control of the retail experience, we developed two platforms in-house: a batch processing platform for processing large amounts of data, and a data hub platform for connecting e-commerce, physical stores, supply chains, and back-office functions. These platforms helped support a seamless connection between customer requests and our manufacturing and delivery capabilities.
Our batch processing platform manages task flow and functions, such as scheduler, distributed processing control, task dependency management, and status management. We sought to simplify the complexity of batch job dependencies to be able to execute parallel distributed processing on a large amount of data within a limited time, and control priority by time zone and the importance of each task.
However, it was difficult to realize these goals with existing third-party products, so we created our own batch processing platform leveraging Amazon Elastic Compute Cloud (Amazon EC2), Amazon DynamoDB, Amazon Simple Queue Service (SQS), Amazon Elastic Container Service (Amazon ECS), Amazon Aurora, and AWS Lambda. The existing third-party product is used only for the scheduler, and batch processing is started by a trigger from the existing commercial product.
Our data hub platform links internal and external systems using this batch processing platform. It is rare that one system, such as the upstream and downstream supply chain and accounting system, consists of only one data channel. Since there are more often many data linkages between systems, we created a hub-and-spoke model Data Hub. Now, inbound processing, routing processing, and outbound processing are defined as task flows, and they are linked in real-time and driven by an event. By relaying with a data hub without connecting with P2P, it is possible to broadcast one inbound to multiple recipients and receive a different protocol and format for each recipient.
In addition to functioning as a layer that absorbs protocol differences, our system also absorbs format differences, such as CSV and XML files. When a system is migrated, the new system and the old system may have different specifications, but our platform lets both operate in parallel without affecting the source system.
As we seek to manifest the future of LifeWear, we wish to create a world-class specialized team. By building a strong customer-driven culture of excellence, our engineers can continually create innovative platforms that serve the unique needs of a business while driving digital transformation in retail.
If you are ready to transform your retail business, then AWS is here to help. Contact your account team today to get started. And if you missed it, check out part 1, where Fast Retailing discusses how they are Transforming the Clothes Shopping Experience.
Fast Retailing Group:
Yasuhiro Kitaguchi, Integration Engineering Team Leader, Digital Business Transformation Services, Fast Retailing Co., Ltd.
Daisuke Hiraoka, Core Engineering Team Manager, Digital Business Transformation Services, Fast Retailing Co., Ltd.
Shimpei Otani, Group Officer and CTO, Digital Business Transformation Services, Fast Retailing Co., Ltd.
Original blog writers from AWS Japan:
Yukako Kunita, Senior Industry SA, Retail
Takuya Kashimura, SA
Takechika Kunimasa, Senior SA