Scholarlink

Among various cloud computing service models, Database-as-a-Service (DBaaS) has become increasingly popular within a multi-tenant architecture. A well-designed Service Level Agreement (SLA) should ensure a balance between maximizing tenant satisfaction, service quality, provider profitability, and minimizing the costs and violations of Service Level Objectives (SLOs). However, achieving all these objectives simultaneously is complex due to workload variation, the multi-tenant environment, and the presence of asymmetric information. In this paper, we propose an Elastic Reservation System (ERS) for a multi-tenant Database Management System (DBMS) in a cloud environment to address the challenges posed by asymmetric information. This system enables the generation of optimal and coordinated SLAs based on a proposed hybrid pricing scheme called the UpdateReservation Scheme (URS). The generated SLA from ERS contains the optimal quantity of reserved resources for each tenant by maximizing the profit functions of both the tenant and the provider. We evaluate the efficiency of this system through simulation-based experiments.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Elastic Reservation System for Multi-tenant DBMSs Through SLA Coordination

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Marzieh Derakhshannia,
  • Franck Morvan,
  • Abdelkader Hameurlain,
  • Oumaima El Haddadi

摘要

Among various cloud computing service models, Database-as-a-Service (DBaaS) has become increasingly popular within a multi-tenant architecture. A well-designed Service Level Agreement (SLA) should ensure a balance between maximizing tenant satisfaction, service quality, provider profitability, and minimizing the costs and violations of Service Level Objectives (SLOs). However, achieving all these objectives simultaneously is complex due to workload variation, the multi-tenant environment, and the presence of asymmetric information. In this paper, we propose an Elastic Reservation System (ERS) for a multi-tenant Database Management System (DBMS) in a cloud environment to address the challenges posed by asymmetric information. This system enables the generation of optimal and coordinated SLAs based on a proposed hybrid pricing scheme called the UpdateReservation Scheme (URS). The generated SLA from ERS contains the optimal quantity of reserved resources for each tenant by maximizing the profit functions of both the tenant and the provider. We evaluate the efficiency of this system through simulation-based experiments.