文章基本信息

标题：Public Clouds Work Sharing Exact Workflows in Time Limits
本地全文：下载
作者：C.Thamizhannai ; S.Nithyadhevi
期刊名称：International Journal of Electronics and Computer Science Engineering
电子版ISSN：2277-1956
出版年度：2014
卷号：3
期号：4
页码：465-468
出版社：Buldanshahr : IJECSE
摘要：An algorithm that uses idle time of provisioned resources and budget surplus to replicate tasks. Deadlines being met and reduces the total execution time of applications as the budget available for replication increases. The description of tasks, data transfer time between tasks if running in different VMs (depicted in the arcs), and execution time of tasks in three different VM types (labeled S1, S2, and S3). The deadline for execution of such workflow is 30 time units and the allocation interval is 10 time units. Deadlines being met and reduces the total execution time of applications as the budget available for replication increases proposed two algorithms for cost-optimized, deadline-constrained execution of workflows in Clouds. On these settings, the IC-PCP algorithm, which is the state-of-the-art algorithm for provisioning and scheduling of workflows in Clouds. THE EIPR ALGORITHM The goal of the proposed Enhanced IC- PCP with Replication (EIPR) algorithm is increasing the likelihood of completing the execution of a scientific workflow application within a user-defined deadline in a public Cloud environment. Task scheduling type of VMs to be used for workflow execution as well as start and finish time of each VM (provisioning). Placement of tasks Data transfer start and end time of scheduled tasks, but also the data transfers to the first scheduled task and from the last scheduled task. Task replication virtual machines to be ready to receive data and tasks in the moment that they are required to meet times estimated during the scheduling process. New criteria for ranking candidate tasks for replication and also workflow structure-aware scheduling of replicas, where the structure of the workflow application is considered not only during the selection of candidates for replication but also during the replica's scheduling. We will also investigate how the replication-based approach can be used when the provisioning and scheduling process is performed for multiple workflows whose requests arrive at different rates.
关键词：Cloud computing; scientific workflows; task replication; soft deadline