CASE STUDY

ADAC

How ADAC's connected vehicle platform was built from scratch, with an MVP delivered within a timeframe that looked impossible.

At a glance

Customer
ADAC
Starting point
Greenfield, no existing platform
Services
Connected Vehicle, Platform Engineering, AI & Data

The outcome

There are no published cost or performance figures for this project. What is on the record is the scope and the pace.

  • MVP delivered in record time, within the original timeframe
  • Greenfield a connected vehicle platform built from scratch
  • 10 years of connected vehicle work behind the delivery

What we did

Context

Since 2016 we have launched a range of connected vehicle solutions, from a complete scooter-sharing platform to event-based processing of IoT device data for both end-user apps and business intelligence. That is the experience ADAC's platform was built on.

Problem

ADAC came to us to build its connected vehicle platform from scratch within a timeframe that looked impossible. There was no existing platform to extend and no time to discover the architecture from first principles.

Approach

The schedule was only achievable because very little had to be invented. We reused what we already knew well: Azure, Terraform, stream processing, Lambda architectures and BI, plus the domain knowledge from earlier connected vehicle projects. Domain knowledge is usually underestimated, and it reduces schedule risk considerably.

Solution

An MVP delivered in record time: vehicle and device events processed as they arrive for the end-user apps, and the same data available for business intelligence, on infrastructure defined as code from the start.

Engineering details

The parts of this project a technical buyer usually asks about.

  • Azure as the platform

    Azure was chosen because the team already knew it well, which made it possible to move quickly within the project's tight timeframe.

  • Terraform for the infrastructure

    Infrastructure defined as code from the beginning, so environments could be stood up repeatedly at the pace the schedule demanded.

  • Stream processing and a Lambda architecture

    Device events are processed as they arrive for the end-user apps, and the same event data feeds batch processing for business intelligence.

  • Domain knowledge as schedule insurance

    Connected vehicle domain knowledge from earlier projects is what made an impossible timeframe achievable, more than any single technology choice.

Need a platform, not another proof of concept?

Tell us what you want your vehicle data to do and what your deadline is. We will come back with the right technical contact and a realistic next step.