An adaptive plan‐oriented and continuous software migration to cloud in dynamic enterprises. (27th June 2019)

• Record Type: Journal Article
• Authors: Nabavi, Seyyed Yahya; Bushehrian, Omid
• Is Part Of: Software, practice & experience. Volume 49:Number 9(2019)
• Page Start: 1365
• Page End: 1378
• Publication Date: 2019-06-27
• Subjects: cloud migration -- finite state process -- migration pattern; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.2725
• Languages: English
• ISSNs: 0038-0644

Abstract

Summary: Replacing existing software/hardware components with their equivalent cloud services is an important decision faced by IT managers in today's enterprises. A variety of possible migration targets and cloud services with too many configurations and cost models, disparate and changing strategic objectives of the enterprise management that triggers the migration process, and the complex structure of the legacy applications make software migration to the cloud a challenging issue. In contrast to the existing approaches that model the migration process as an optimization problem to find the optimal deployment of software components on cloud services without presenting a practical migration plan, in this paper, a plan‐oriented migration approach is proposed by which the enterprise management is able to follow migration steps of a valid plan. All valid plans are modeled using a labeled transition system, and a recommender engine directs the management through the possible migration paths using predefined fitness functions. It was observed that, particularly in dynamic and changing conditions that a flexible migration plan is essential, the proposed plan‐oriented method is very much effective in satisfying the enterprise strategic objectives. Evaluations have been performed using two quality indicators: total cost of ownership and scalability index.