Players such as Google, Amazon, or Microsoft offer a plethora of alternative cloud-based services. The decision to move from legacy IT infrastructure or existing cloud-based services to another cloud-based solution for fulfilling new or existing IT requirements in an optimized manner for any organization is not a trivial process. Such a decision is affected not only by different alternative solutions, but also by contradictory or mutually dependent influencing criteria or factors. Therefore, decisions to adopt cloud-based services or any new technology better follow a quantified trade-offs based methodology.

Today, the decision-making method in organizations for the selection of cloud-based solutions is an ad-hoc process that is solely based on the market reputation of the Cloud Service Provider and past experiences of IT (Information Technology) decision makers within an organization. Even though, these are important factors in such a process, they are by far not sufficient as such decisions do not have an objective and quantitative basis. Such a decision-making process of selecting the best alternative falls within the category of Multiple Criteria Decision Analysis (MCDA). One of the MCDA algorithm namely Analytical Hierarchy Process (AHP) has been used to rank cloud-based services by structuring relevant factors in a hierarchy. However, the problem is that this approach still lacks a holistic view of integrated relevant factors from technical, economical, and organizational domains and their interrelations forming a complex network of many inter-dependent or even conflicting factors.

TrAdeCIS is evaluated with respect to four use cases that involved decisions of adopting cloud-based services of organizations who participated in exploratory research. These evaluations varied in complexity of the decision models due to the inclusion of different relevant factors, their interrelations, and alternatives. It concluded that TrAdeCIS can be applied to model and make such quantitative decisions. Performance evaluation of TrAdeCIS calculated the execution time of ranking 100 alternatives using 100 technical criteria, and 100 economical and organization criteria (which is the upper limit of factors based on the exploratory research) as fast as below 20 ms. This is achieved an optimized implementation of TrAdeCIS to ensure that results obtained are not outdated due to dynamically changing input in performance values of an alternative. An application of TrAdeCIS, for extensibility and generalization evaluations, to the decision of choosing the best technology by train operating companies to improve both voice and data coverage on-board of trains proves that TrAdeCIS is even valid for a decision of adopting any new technology in an organization. Hence this thesis concludes that the only requirement for a general applicability of TrAdeCIS is that the decision must involve multiple alternative solutions, which have to be evaluated for multiple criteria.

Hence, this thesis investigates the following three aspects that provide for seamless unification of multiple alternatives, relevant factors, and their interrelations under a single developed system for decision making of adopting a new technology, specifically Cloud Computing. Firstly, exploratory research led to the development of a new structured taxonomy consisting of 102 factors and their interrelations. This coherent taxonomy forms the basis for evaluating the performance of alternative cloud-based services from all relevant perspectives. Secondly, a Trade-offs Based Methodology of Adopting Cloud-based Services (TrAdeCIS) is designed and implemented prototypically as a Web-based platform for ranking alternatives. This performs and supports quantified trade-offs-based decisions for selecting the best technical and at a trade-off of business value, if the ranking of alternatives from technical and business perspective is not the same. This platform also facilitates demonstrations and use-case based evaluations of TrAdeCIS. Thirdly, a predictive Impact Analysis Methodology for Cloud-based Services (IAMCIS) is developed to measure the impact of adopting the top ranked alternative as per TrAdeCIS. This leads to the identification of potential risks associated to any failure of services.

Schlagwörter: Computer Science; Decision Analytics; Operations Research; Cloud-based Services; Multi-attribute Decision Algorithm; Cloud Computing

Informatik