1
 
 
Account
In your account you can view the status of your application, save incomplete applications and view current news and events
OTTO goes AWS - Part 1
October 04, 2018

OTTO goes AWS - Part 1

160+ Services, 24 Teams, 1 year

The cloud stands for scalability, cost efficiency and short time to market for new features. otto.de also wants to benefit from this and therefore started the migration to AWS last summer.

The first question was how the whole thing should be organized. Centralized or decentralized? Agile or conservative project management? We decided on everything at the same time - a largely decentralized implementation with a hybrid project management approach.

Organizationally, a central migration team was created to support our decentralized development teams. The motto here is always help for self-help by means of a mentoring conceptEach team therefore had a fixed contact person with whom architecture decisions, concrete implementation details or questions relating to the cultural change were discussed. An accompanying framework program consisting of six workshops was also established, in which central concepts, problems or call-for-support were discussed from team to team. The agenda was generally designed in an agile manner and changed several times a day as needed.

The development teams continued to work in agile mode and regularly synchronized progress with the central project plan, which was a kind of milestone plan. Risks could thus be identified at an early stage and minimized by taking appropriate measures.

Important issues during the migration were

  • Central vs. decentralized implementation of certain aspects
  • Strategy in supporting our development teams
  • Implementation of a security strategy
  • Communication between our individual services, esp. VPN or no VPN?
  • Dealing with vendor lock-in effects
  • In-house development of services vs. standardized high-level services

These and other questions will be considered in subsequent articles.

But what was the biggest challenge? It was not the technological implementation, but rather the cultural shift toward decentralized operation of the infrastructure. Previously, operational responsibility rested with a central operations unit. Now the teams are completely autonomous and have to deal with issues such as readiness or disaster recovery. So it makes sense in any case to accompany such a project with appropriate change management measures.

After this very exciting and eventful year, the foundation has now been laid for the ever-growing requirements - e.g. the platform project.

0No comments yet.
Write a comment
Leave us a comment here and let the authors know what you thought of the article.
Answer to: Reply directly to the topic

Written by

Similar Articles

Saved!

We want to improve out content with your feedback.

How interesting is this blogpost?

We have received your feedback.