• analysis
• container orchestrators
• deployment
• microservices

The overall objective of the thesis is to support an effective analysis of microservice-based applications and the automation of their deployment over container-based platforms. The main contributions of the thesis are (i) a methodology for automating the detection of architectural smells possibly violating some main principle of microservices, and for resolving such smells via architectural refactorings, (ii) a technique for completing an application specification by automatically discovering Docker-based runtime environments capable of supporting the application components, and (iii) an architectural approach for effectively deploying microservice-based applications on top of existing container orchestrators, by also allowing to manage the lifecycle of microservices independently of the lifecycle of the containers hosting them. All the aforementioned solutions have been implemented into running prototypes and tested on concrete case studies, involving third-party software solutions.