A Devops Platform For Microservices

Bose Corporation started producing Internet-connected speakers and, therefore, needed to build out a previously non-existing software development practice in the cloud, having historically been focused primarily on acoustic and audio technology. Our consulting services have been pivotal in Bose’ adoption of the cloud and still support their maturation curve.

Initially, Bose Corporation utilized aimtheory’s professional services to modernize their infrastructure automation and add continuous delivery capabilities. We used a combination of Jenkins, Chef and AWS Cloudformation as the core solution to automatically test new code commits, provision cloud resources in Amazon Web Services and then deploy applications–like Bose’ Amazon Alexa integration–to those resources. During this time we also added various other capabilities around monitoring, metrics, enterprise authentication integration and logging with tools like logstash, ElasticSearch and DataDog.

As Bose began to execute their plan to create a custom, in-house IoT platform that would containerize and deliver all of their future ecosystem microservices, they once again utilized aimtheory’s professional services to build out the first generation of this platform. During this time we helped architect and deliver an 0.1 release of the platform to production. Heavily focused on Cloud Native Computing Foundation projects, we helped evaluate and select Kubernetes as the backbone of the project. We worked with other teams to lead the design and build of custom Jenkins pipelines, a custom developer portal and various custom tools in Python, Ruby and others. All of these tools would come together to provide a developer experience that would allow them to quickly and easily define their entire microservice and delivery pipeline–from test automation, to continuous delivery process to automated deployments in a unified markup file. The platform is polyglot, supporting the continuous delivery of microservices in many languages including Java, Python, Ruby and NodeJS. The platform is in production and continues to evolve from this initial release as it supports millions of devices.

• Client: Bose Corporation
• Date: December 2016