Dynamic microservices placement in multi-cluster infrastructures

During the composition of my Thesis Work, I engaged with the challenge of Optimizing Response Times in a Multi-Cluster Kubernetes Infrastructure. The infrastructure had a three-layer topology, incorporating Cloud, Fog, and Edge computing paradigms. Placing microservices closer to the Edge is deemed to reduce the response time of an application as opposed to services placed at the upper layers. However, Fog and Edge nodes are characterized by their limited resources. Thus choosing among microservices to run on the Edge or the Fog clusters was the problem my Thesis was dealing with. On the premises of my Thesis, I developed an application, namely DeFog. DeFog focuses on dynamic decentralized microservice placement within each cluster, eliminating the need for coordination between them. With the help of Linkerd Service Mesh, DeFog gathers essential data and monitor applications. Utilizing these technologies, DeFog makes informed decisions for microservice placement, all while accomplishing efficient load balancing.

Link for my Thesis:


Link for medium article I posted:

That’s very cool @aprountzos . Thanks for sharing!

1 Like