Shared Resource-Aware VM Scheduling in the Cloud

Description

Commodity  servers deployed in the data centers are now typically using the  Non-Uniform Memory Access (NUMA) architecture. The NUMA multicore  servers provide scalable system performance and cost-effective property.  However, virtual machines (VMs) running on NUMA systems will access  remote memory and contend for shared on-chip resources, which will  decrease the overall performance of VMs and reduce the efficiency,  fairness, and QoS that a virtualized system is capable to provide. In  this project, we design and implement a shared resource-aware VM  scheduler on NUMA multicore servers that are now commonly used in the  cloud.

Shared Resource Usage Analysis

VM Performance Modeling and Prediction

Architecture Overview

Online VM Performance Monitoring

Shared Resource-Aware VM Scheduler on KVM Platform

Shared Resource-Aware VM Scheduler on Xen Platform

Future Work

  • Improving VM Performance and Server Utilization to Reduce Cost in the Cloud

  • Improving VM QoS guarantee in the multi-tenant cloud environment

  • An intelligent VM management system that provides both efficiency and fairness for the cloud users

Members

  •     

Copyright Statement

This  material is presented to ensure timely dissemination of scholarly and  technical work. Copyright and all rights therein are retained by authors  or by other copyright holders. All persons copying this information are  expected to adhere to the terms and constraints invoked by each  author's copyright. In most cases, these works may not be reposted  without the explicit permission of the copyright holder.


<<< 返回