基于嵌入式虚拟化平台的域间数据共享系统
发布时间:2018-07-18 20:48
【摘要】:随着嵌入式设备的迅猛发展,嵌入式虚拟化技术也成为工业界和学术界关注的热点。虚拟化技术在资源整合以及安全隔离等方面的优势非常突出。然而,虚拟机天然的隔离特性使得虚拟机之间的数据共享变得相对困难。目前,虚拟机之间通常利用网络实现数据共享,例如Network File System、Samba等,但是,这种方法不能保障数据传输的高可靠和低延迟,不符合嵌入式系统对可靠性、实时性和能耗的要求。 嵌入式虚拟化域间数据共享系统有效的解决了上述问题。该系统的特色是:其域间数据共享技术完全不依赖网络,而是充分利用了VNeo嵌入式虚拟化系统的设备模型,将共享块设备驱动逻辑放在嵌入式虚拟机的管理器中实现,虚拟机管理器根据块设备驱动逻辑操纵硬件,完成数据读写。在客户操作系统中,用户态文件读写进程通过系统调用向内核发出读写请求。内核负责处理读写请求信息,,并通过超级系统调用将文件读写请求信息发送给虚拟机管理器。虚拟机管理器根据共享设备的驱动逻辑驱动设备进行文件读写。数据读写完毕后,虚拟机管理器通过虚拟中断的方式通知客户操作系统数据读写完成。客户操作系统的虚拟中断处理程序响应虚拟中断,唤醒阻塞的读写进程。 系统测试与结果分析表明:采用该数据共享系统对共享块设备的文件进行读操作的性能达到了未虚拟化时读取本地块设备上的文件性能的90%,写性能达到了未虚拟化时写本地块设备性能的70%。该系统符合嵌入式系统所要求的高可靠、低延迟和低能耗等特点。
[Abstract]:With the rapid development of embedded devices, embedded virtualization technology has become a hot spot in industry and academia. Virtualization technology in the integration of resources and security isolation and other aspects of the advantages are very prominent. However, the natural isolation of virtual machines makes it relatively difficult to share data between virtual machines. At present, network is used to share data between virtual machines, such as Network File system, Samba, etc. However, this method can not guarantee high reliability and low delay of data transmission, and does not meet the requirements of embedded system for reliability, real-time and energy consumption. Embedded virtualization inter-domain data sharing system effectively solves the above problems. The characteristic of the system is that the technology of data sharing between domains is completely independent of the network, and it makes full use of the device model of VNeo embedded virtualization system, and implements the shared block device driver logic in the manager of embedded virtual machine. The virtual machine manager manipulates the hardware according to the block device driver logic and completes the data reading and writing. In the client operating system, the user state file read and write process sends the read and write request to the kernel through the system call. The kernel is responsible for handling the read and write request information, and sends the file read and write request information to the virtual machine manager through the super system call. The virtual machine manager reads and writes files according to the shared device's driver logic device. After the data read and write, the virtual machine manager notifies the client operating system to complete the data reading and writing by the way of virtual interrupt. The client operating system's virtual interrupt handler responds to the virtual interrupt and wakes up the blocked read and write process. The system test and result analysis show that the performance of the data sharing system to read the files of the shared block device has reached 90% of the performance of reading the files on the local block device when it is not virtualized, and the performance of writing has reached the level of non-virtual. 70 of the performance of the piece of land equipment. The system meets the requirements of embedded system such as high reliability, low delay and low energy consumption.
【学位授予单位】:华中科技大学
【学位级别】:硕士
【学位授予年份】:2012
【分类号】:TP368.1
本文编号:2132731
[Abstract]:With the rapid development of embedded devices, embedded virtualization technology has become a hot spot in industry and academia. Virtualization technology in the integration of resources and security isolation and other aspects of the advantages are very prominent. However, the natural isolation of virtual machines makes it relatively difficult to share data between virtual machines. At present, network is used to share data between virtual machines, such as Network File system, Samba, etc. However, this method can not guarantee high reliability and low delay of data transmission, and does not meet the requirements of embedded system for reliability, real-time and energy consumption. Embedded virtualization inter-domain data sharing system effectively solves the above problems. The characteristic of the system is that the technology of data sharing between domains is completely independent of the network, and it makes full use of the device model of VNeo embedded virtualization system, and implements the shared block device driver logic in the manager of embedded virtual machine. The virtual machine manager manipulates the hardware according to the block device driver logic and completes the data reading and writing. In the client operating system, the user state file read and write process sends the read and write request to the kernel through the system call. The kernel is responsible for handling the read and write request information, and sends the file read and write request information to the virtual machine manager through the super system call. The virtual machine manager reads and writes files according to the shared device's driver logic device. After the data read and write, the virtual machine manager notifies the client operating system to complete the data reading and writing by the way of virtual interrupt. The client operating system's virtual interrupt handler responds to the virtual interrupt and wakes up the blocked read and write process. The system test and result analysis show that the performance of the data sharing system to read the files of the shared block device has reached 90% of the performance of reading the files on the local block device when it is not virtualized, and the performance of writing has reached the level of non-virtual. 70 of the performance of the piece of land equipment. The system meets the requirements of embedded system such as high reliability, low delay and low energy consumption.
【学位授予单位】:华中科技大学
【学位级别】:硕士
【学位授予年份】:2012
【分类号】:TP368.1
【参考文献】
相关期刊论文 前8条
1 左源;刘新宁;师超;;一种SD卡控制器的硬件实现[J];电子器件;2007年02期
2 徐洋;史学书;刘明波;;基于PCI核的链式DMA的设计[J];国外电子测量技术;2008年06期
3 韩辉;丛培田;郭颖;;基于SD卡的多通道数据采集系统[J];工业仪表与自动化装置;2010年01期
4 刘真;;虚拟机技术的复兴[J];计算机工程与科学;2008年02期
5 邹念,唐宁九,林锋;用Samba实现Linux和Windows之间的文件共享[J];计算机应用研究;2002年01期
6 刘欣;郑建宏;;基于ARM9的SD/MMC卡控制器的ASIC设计[J];通信技术;2008年08期
7 汤龙梅;张国安;;基于S3C2410的SD/MMC卡驱动的实现[J];微计算机信息;2009年17期
8 吉涛;蔡航;;应用SD卡实现大容量存储测试系统设计[J];陕西科技大学学报(自然科学版);2010年05期
本文编号:2132731
本文链接:https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/2132731.html