Date of Award

Spring 3-2022

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Business and Information Systems

First Advisor

Dr. Omar El-Gayar

Second Advisor

Dr. David Bishop

Third Advisor

Dr. Yong Wang

Abstract

A recent trend in software development adopts the paradigm of distributed microservices architecture (MA). Kubernetes, a container-based virtualization platform, has become a de facto environment in which to run MA applications. Organizations may choose to run microservices at several cloud providers to optimize cost and satisfy security concerns. This leads to increased complexity, due to the need to observe the performance characteristics of distributed MA systems.

Following a decision guidance models (DGM) approach, this research proposes a decentralized and scalable framework to monitor containerized microservices that run on same or distributed Kubernetes clusters. The framework introduces efficient techniques to gather, distribute, and analyze the observed runtime telemetry data. It offers extensible and cloud-agnostic modules that can exchange data by using a multiplexing, reactive, and non-blocking data streaming approach. An experiment to observe samples of microservices deployed across different cloud platforms was used as a method to evaluate the efficacy and usefulness of the framework. The proposed framework suggests an innovative approach to the development and operations (DevOps) practitioners to observe services across different Kubernetes platforms. It could also serve as a reference architecture for researchers to guide further design options and analysis techniques.

Share

COinS