The Project
Together with public transport, TIER offers emission-free alternatives to car rides, ultimately freeing cities from pollution and congestion. TIER’s e-scooters are rented in a free-flow model by just scanning a qr code - combined with a seamless registration model, the customer journey from registration to first ride can be completed in well under one minute. Several integrations into several popular public transportation apps such as e.g. the BVG Jelbi App for Berlin and the MVG More App for Munich have been carried out.
The Challenge
When starting up, TIER faced an extremely competitive environment combined with high customer expectations for differentiation. This led to a number of engineering decisions by the company’s leading technologists CTO Matthias Laug and Directory of Technology Amir Hadi: first and foremost they wanted simple and uniform deployment interfaces with the least amount of operations upkeep possible for their engineers. The second pillar was a full-automation approach in order to make infrastructural decisions reviewable and refactorable, and any build and subsequent deployment processes seamless and simple. The last pillar was going to be a datalake driven business analytics infrastructure, collecting sensor data, business transaction metrics, access logs and other forms of telemetry while baking in GDPR compliance requirements from the very beginning.
“Our plan was to make it a no brainer to deploy new internal and customer-facing features fearlessly while also make it trivial for engineers and business analysts alike to measure it’s impact,” explains Amir Hadi.
In order to implement these organisational requirements without infringing on the freedom of developers, TIER built their architectures implicitly around the pillars of the well-architected framework. This means that operational excellency (specifically a best-of-breed, end-to-end monitoring), security, reliability, performance efficiency and cost were first-class objectives when developing their delivery. Even in production and accelerating its operations, TIER today provides freedom for its developers to implement their solutions in an optimal solution space: from Spring Boot Java applications to lean & mean Node applications and statically linked Golang binaries.
The Solution
Best of breed deployment interfaces and deep datalakes
AWS turned out to be the ideal infrastructure partner for almost all of these requirements. Supported by kreuzwerker, TIER automated the infrastructure of its two primary deployment interfaces: AWS Lambda and AWS Elastic Kubernetes Services. While nominally covering the same space, the usage of both Lambda and K8S underlines the best of breed approach pursued by TIER: Lambda today handles most of the event-driven, short lived business intelligence related tasks, as well as some APIs build using a large number of highly decoupled function written using NodeJS functions and orchestrated through the serverless framework which - in turn - uses the AWS CloudFormation infrastructure provisioner. The elastic K8S services on the other hand tend to run the compute heavy parts of the infrastructure, and are supported by some AWS managed services such as RDS / Aurora. The same stack supports business analytics toolchains running data transformation and query services on top of S3 sourcing databases such Athena and AWS Redshift. This infrastructure is wholly automated in the same fashion using the Terraform infrastructure provisioner built by AWS Technology Partner HashiCorp.
Continuous integration and deployment services running a mixture of Helm, a custom Terraform Enterprise-style setup and serverless deployments are handled by AWS Technology Partner CircleCI, while cross-stack infrastructure, K8S and log monitoring services are provided by the AWS Technology Partner DataDog with additional help by cloud-native heavyweights Grafana and Prometheus.
The Upshot
Next steps for TIER and the future of mobility
Today TIER processes multiple TB of events every week generated by 750,000 rides performed by thousands of riders and is continuously expanding its operations into new cities and markets. Tomorrow it’ll move down the layers inside their IoT-enabled next-generation e-scooter adding AWS managed services on the way. At the same time, TIER is aggressively adapting the datalake paradigms of AWS embracing technologies such as Lake Formation, AWS Glue and AWS Athena to drive its next generation business intelligence operations. This parallel strategy of unlocking more telemetry without compromising privacy expectations, while extending the depth of insights into this data, will ultimately lead to better insights, a better product, happier riders and a greener city.
“Our new e-scooters will continuously improve in comfort and will combine a great driving experience with sustainable and emission-free mobility,” says Lawrence Leuschner, CEO of TIER.
AWS’ strong technical foundation and higher level services will continue to support this vision going forward.