Clustering-Based Cloud Migration Strategies

Abstract

The k-means algorithm of the partitioning clustering method is used to analyze cloud migration strategies in this study. The extent of assistance required to be provided to organizations while working on migration strategies was investigated for each cloud service model and concrete clusters were formed. This investigation is intended to aid cloud consumers in selecting their required cloud migration strategy. It is not easy for businessmen to select the most appropriate cloud migration strategy, and therefore, we proposed a suitable model to solve this problem. This model comprises a web of migration strategies, which provides an unambiguous visualization of the selected migration strategy. The cloud migration strategy targets the technical aspects linked with cloud facilities and measures the critical realization factors for cloud acceptance. Based on similar features, a correlation among the migration strategies is suggested, and three main clusters are formed accordingly. This helps to link the cloud migration strategies across the cloud service models (software as a service, platform as a service, and infrastructure as a service). This correlation was justified using the digital logic approach. This study is useful for the academia and industry as the proposed migration strategy selection process aids cloud consumers in efficiently selecting a cloud migration strategy for their legacy applications.

Cloud migration framework clustering method for social decision support in modernizing the legacy system

“…Additional details on cloud migration can be found in our article. 19 The benefits of cloud computing prompt the migration of legacy systems to a cloud environment. 5,23 Service-oriented environments aid in improving the ability to migrate a legacy system as a service in a service-oriented architecture (SOA).…”

“…A clustering technique is different from a classification technique. 19 Classification, being a form of supervised learning, analyzes the classes labeled or instances that are already defined. In contrast, clustering is not involved in such training and is usually performed when the sets are not defined.…”

“…40 Therefore, the clustering technique can assist in grouping the cloud stuff such as cloud migration frameworks. 19 The relationships among clusters can be identified by various measurement techniques, 41 but two statistical measures in particular, called distance measure and similarity measure, 42 are commonly used. In Reference 19, the similarity was measured using metrics such as the Jaccard coefficient, 15 whereas, in the present study, K3CM measurements are proposed.…”

“…The details on the cloud migration strategy selection process can be taken from our article. 19 The state-of-the-art cloud migration framework for legacy systems is ADM, and it was chosen for the present study. As technology advances, any new modernization framework could replace ADM. Other cloud migration frameworks selected from the literature review were compared with ADM.…”

“…There is a need for a process by which the industry can easily decide and select the best cloud migration frameworks for their legacy system. 10,19 Following the proposed process, selected features of ADM and other cloud migration frameworks were compared to check their similarity (to identify the most commonly used features for successful cloud migration).…”