Communication-aware management of SLAs for Cloud-Native Applications.On the road to AIOps. Smart orchestration strategies in Cloud and Edge computing.

Today, IT Operations teams have to face up with managing massive amounts of data generated by advanced distributed systems, workloads difficult to predict in time, security threats. They need to handle more incidents than ever before with strict service-level agreements (SLAs). Most of the current state-of-the-art techniques to handle SLAs for Cloud-Native applications are based mostly on severe human efforts. Downtime can get expensive: companies can lose millions dollars per outage with a longer mean time to recovery due to the complexity of human debugging on complex distributed systems. In the landscape of hybrid clouds, multi-tenant environments, and Edge computing architecture, organizations need multiple strategies to get the desired quality of service. Capacity planning, resource utilization, storage management, anomaly detection, threat detection are just a few aspects that engineering teams should take into account to guarantee SLAs and sites reliability. AIOps can empower software and service engineers to effectively build and operate cloud-native applications at scale with artificial intelligence (AI) and machine learning (ML) techniques. This thesis explores the multitude of issues that constitute the landscape of cloud-native applications in Cloud and Edge computing scenarios. The focus will be twofold: proposing containers allocation strategies where communication quality is considered first-citizen parameter, and opening a window in the novel field of AIOps by providing models, case studies, and strategies to perform smart orchestration of run-time workloads in PaaS clouds.