Introspection of Virtual Machine Memory Resource in the Virtualized Systems

Nowadays, data center and cloud servers have commonly adopted virtualization technologies to consolidate multiple servers into a physical one. It is because cloud service providers can achieve low server maintenance cost by improving resource utilization and reducing power consumption through server...

Full description

Saved in:
Bibliographic Details
Main Authors: Lee, Minho, Park, Sujin, Song, Yongju, Eom, Young Ik
Format: Conference Proceeding
Language:English
Subjects:
content-based memory sharing
Kernel
Memory management
Monitoring
Servers
virtual machine migration
Virtual machining
virtual-machine monitoring tool
Virtualization
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Nowadays, data center and cloud servers have commonly adopted virtualization technologies to consolidate multiple servers into a physical one. It is because cloud service providers can achieve low server maintenance cost by improving resource utilization and reducing power consumption through server consolidation with virtualization technologies. However, unlike the physical system resources such as CPU and storage which can be flexibly utilized and shared by using time-based schedulers, memory resource is not easy for flexible utilization and sharing in that the memory size of each virtual machine is fixed by initial configuration. For this reason, sufficient understanding on memory resource usage of each virtual machine is essential in analyzing the existing memory management techniques such as memory ballooning and virtual machine migration. In this paper, we introduce a novel virtual machine memory monitoring tool, called SELF-e, which is developed for tracing the page accesses of each virtual machine in real-time and collecting necessary information on shared pages. Experimental results show that SELF-e efficiently announces the information on classified pages without significant performance degradation.
ISSN:2375-9356
DOI:10.1109/BIGCOMP.2019.8679210