ifgerman{chapter*{Inhaltsangabe}}{chapter*{1.Introduction}}Nowadays, lack of enough human resources as well as

ifgerman{chapter*{Inhaltsangabe}}{chapter*{1.Introduction}}Nowadays, the cloud computing technology has fundamentally transformed the way applications developed.citep{xiao2013security}.A few years ago, Companies were enforced to have dedicated teams for configuring, running and maintaining server rooms which made it extremely difficult to scale up easily and offer a sustainable product to users. It was even more difficult for small startup companies due to the lack of enough human resources as well as funding.This is where cloud computing has come into the picture and changed the ways in which software is developed, operated and consumed.citep{armbrust2010view}.virtualization techniques are the backbone where this cloud computing relies on to achieve elasticity of large-scale shared resources.Virtual machines(VMs) and containers are vital virtualization techniques to provide virtualized operating systems (OSs) at the infrastructure layer.Despite virtual machines, the containers are not only more light-weight virtualization concept but also they are less resource and time-consumingcitep{pahl2015containerization}. A container allows the operator to run application code, required libraries and the binaries, and a basic configuration which is needed to run the applicationcitep{casalicchio2016autonomic}.These containers are very efficient and also many of them can run on the same machine, allowing full utilization of all available resources. The cross-contamination between containers can be avoided by Linux containers and Cgroups.Libraries, ports, namespaces, and memory contents are all kept isolated. They also enforce upper boundaries on how much system resource (memory, storage, CPU, network bandwidth, and disk I/O) a container can consume so that a critical application isn’t squeezed out by noisy neighbours.Docker is the most powerful container technology with its ease of use and allows developers and operators to easily deploy their applications in containers. Docker and containers became so popular because they potentially can solve multiple cloud application problems like “dependency hell” problem, application portability problem and performance overhead problem citep{casalicchio2016autonomic}.These container runtime API’s can manage one container on one host only. They are not suited to manage multiple containers on multiple hosts. In enterprises, We need to run multiple containers in the multi-host environment. This is where we need to look at container orchestration tools as they offer distinct plugins to overcome the above problem.subsection{Problem Identification and Motivation}   There are several container orchestration tools available in the market. It is always challenging task to enterprises in selecting a consistent orchestration tool for respective cloud environment while keeping the focus on what is important and valuable for the enterprise. As no single container orchestration tool has the perfect solution for the unique business situation, the only way to maximize the benefits of container orchestration tool services to enterprises is by dealing with the performance of several container orchestration tools and monitoring all of them simultaneously. As a fast scaling startup, Wunder carpool has been using various public cloud services like Amazon web services 3, Google cloud 4 and DigitalOcean 5 to provide carpooling services to the customers. This research would also consider another popular public cloud service like Microsoft Azure 6 for the future prospects. This research will investigate container orchestration in public cloud environments and explore to establish a consistent orchestration tool for Docker 7 in various public clouds or to use respective container orchestration solutions for the respective cloud. This research will use the following orchestration tools to examine and evaluate with the specific set of features: Kubernetes 8, Docker Swarm 9 and Mesosphere 10.                    In the end, this master thesis will develop a decision-making table with the research results and concluding that which container framework is best fit on multi cloud   environment for a small size organization like wunder carpool.