报告题目：Cloud YARN resource provisioning for task-batch based workflows with deadlines.
摘要：To meet the dynamic workload requirements in widespread task-batch based workflow applications, we developed the cloud YARN (C-YARN) architecture by integrating the YARN platform with cloud computing. The C-YARN could provision flexible resources. In terms of depths and functions, tasks of different task-batches were merged to task-units. Based on task-units, a unit aware deadline division method was investigated for properly dividing workflow deadlines to task deadlines to minimize the utilization of rented intervals. Considering different factors affecting time slot allocation, several rules were introduced for scheduling tasks with the task deadlines. A rule-based task scheduling method was presented for allocating tasks to time slots of rented Virtual Machines with a task right shifting operation and a weighted priority composite rule.
A Unit-aware Rule-based Heuristic was proposed for elastically provisioning VMs to task-batch based workflows to minimize the rental cost in C-YARN. Effectiveness of the proposals was verified by comparing them against two adapted existing algorithms for similar problems on some realistic workflows.
李小平教授简历: 李小平，男，教授，博士生导师，工学博士，IEEE高级会员，CCF高级会员。主持国家863计划、国家自然科学基金等10余项科研项目；在IEEE Transactions on Automation Science and Engineering, IEEE Transactions on Services Computing, Omega, European Journal of Operational Research, Information Sciences, ICSOC, ICPADS等国际期刊或会议上发表论文100余篇。2012入选江苏省“六大人才高峰” 培养对象，2009年入选教育部新世纪优秀人才支持计划。主要研究内容：调度理论与算法、服务计算、企业互操作技术、云服务供应方法。