Update of software usually equates disruption to a program’s execution. However, such disruptions in the execution of applications that provide round-the-clock services is not desirable. Examples of such applications include multi-tier enterprise systems with which users may interact via distinct presentation tiers at the same time. Existing software updating systems do not operate in the enterprise applications, or they need most of the redeployment tasks to be performed manually. In this paper, we present a framework capable of Automatically Redeploying Enterprise Software Applications (ARESA). This system combines the idea of incremental and integral updates to preserve the consistency of applications and to minimize disruption times during redeployments at the same time. It also utilizes the update bringing forward techniques in order to enhance the system’s predictability. ARESA has been used in updating a desktop costing application, a custom web program, and Joomla content management system. The experimental results show that in applications with real-life complexities, ‘speed’ and ‘size’ overheads incurred by ARESA are less than 19% and 1%, respectively.
Digital Object Identifier (DOI)
Seifzadeh, Habib; Abolhassani, Hassan; and Sadighi Moshkenani, Mohsen
"A Performance-based Approach to Automatic Redeployment of Enterprise Software Applications,"
Applied Mathematics & Information Sciences: Vol. 10
, Article 15.
Available at: https://dc.naturalspublishing.com/amis/vol10/iss2/15