2020 International Conference on Advance in Ambient Computing and Intelligence (ICAACI)

Design and Implementation of Operating System in

Distributed Computer System Based on Virtual
2020 International Conference on Advance in Ambient Computing and Intelligence (ICAACI) | 978-1-7281-9753-1/20/$31.00 ©2020 IEEE | DOI: 10.1109/ICAACI50733.2020.00024

Machine

Xing Rong
University of Miami
Miami, Florida, USA

Abstract—Distributed computer network system is a resources [5].

network system with high cohesion, transparency and
autonomy. Users of a distributed system see a single, high- A lot of experiments and work have shown that the
performance and high-reliability data computing and virtual machine operating system is actually a very time-
processing system with a large number of resources. With the consuming process in the specific deployment process [6].
long-term development of virtualization technology, the This time-consuming can actually be effectively answered
performance of virtual machine can meet many application in several aspects such as its large mirror image and poor
requirements, especially for the virtualization of processor and storage methods. In order to overcome the communication
memory, which has reached the performance similar to that of bottleneck in the multi-processor system, people began to
physical host. With the popularity of desktop virtualization use networking technology to construct a multi-computer
technology, many enterprises begin to deploy desktop system, that is, a distributed computer system [7]. For a
virtualization system. Nowadays, the virtual machine with distributed system, designing a DNOS that can meet the
certain characteristics and performance advantages has system target requirements and application environment
gradually matured after a long time of development, and has requirements is the key to the realization of the distributed
been widely used in computer related industries. Based on the system [8]. However, DNOS itself is a large-scale software
connotation and characteristics of virtual machine, this paper subject with great difficulty and long development cycle. It
analyzes and studies the design and implementation process of requires good software design ideas and methods as a basis
operating system in distributed computer system.
to guide the design and implementation of the system. Due
Keywords—Distributed computer, Virtual machine,
to the limited throughput rate of a single storage service,
Operating system such virtual desktops can often only support dozens of
virtual machines in one storage view, which cannot meet the
I. INTRODUCTION requirements of desktop virtualization deployment in
medium and large enterprises [9]. The use of a distributed
Distributed computer network system, referred to as
storage structure can effectively improve the throughput
distributed system, is a network system with high cohesion,
capacity of the storage system. In the metadata management,
transparency and autonomy [1]. With the popularization of
the direct block index structure is used to obtain better
desktop virtualization technology, many enterprises have
snapshot performance, and the redundancy technique is used
begun to deploy desktop virtualization systems. Nowadays,
to reduce the redundant data between multiple virtual
virtual machines with certain characteristics and
machine mirrors [10]. Based on the connotation and
performance advantages have gradually matured after a long
characteristics of virtual machines, this paper analyzes and
period of development, and have been widely used in
studies the process of operating system design and
computer-related industries. The realization of this
implementation in distributed computer systems.
widespread use actually depends to a large extent on its
deployment efficiency [2]. After long-term development of II. ANALYSIS OF THE CONNOTATION AND CHARACTERISTICS
virtualization technology, its virtual machine performance OF VIRTUAL MACHINE
can already meet the needs of many applications, especially
for the virtualized processing of processors and memory, Virtual machine is a virtual simulated computer system,
which has reached performance similar to that of physical which can realize the corresponding hardware and software
hosts [3]. In order to improve the computing power of the functions by installing and running the corresponding
system, people have great expectations for the idea of using software. In the design of DNOS, it is not easy to achieve
multiple CPUs to construct a new system. The development these goals truly and uniformly, and a good system design
of large-scale integrated circuits and microcomputer method is needed. On the same computer, multiple virtual
technology provides a realistic possibility for the realization clients can be set up at the same time, and several virtual
of multiprocessor systems [4]. Distributed computer systems operating systems can be run simultaneously among the
have been greatly developed in recent years, and have running virtual clients [11]. In the computer network, each
received more and more attention from the scientific and computer has its own operating system, the resources are
technological circles, and their applications have become locally owned and controlled, and the process scheduling in
more and more attractive. What a distributed system user the network is realized by migrating processes and data. In
sees is a single, high-performance, and highly reliable data the distributed computer system, a part of the global and
computing and processing system with a large number of system-wide operating system runs on each node computer.

978-1-7281-9753-1/20/$31.00 ©2020 IEEE 94

DOI 10.1109/ICAACI50733.2020.00024

Authorized licensed use limited to: Miami University Libraries. Downloaded on June 14,2021 at 20:14:37 UTC from IEEE Xplore. Restrictions apply.
As far as the virtual machine is concerned, it always needs remote startup technology and installation image creation
the deployment of its virtual machine system as an effective technology. If there is interaction among all parts, it is
premise, because as far as this virtual system is concerned, it accomplished through semantic-based information
mainly includes the existing operating system and the brand- transmission. In this case, in order to ensure the correctness
new virtual system. As far as this kind of virtual machine is and reliability of the whole system, it is only necessary to
concerned, it is quite different from the virtual system, ensure the correctness and reliability of the interaction
because the virtual machine usually needs to simulate the between each part and each part. In a distributed operating
corresponding hardware instructions in the running process, system, resources are managed and used globally, and their
and the running speed is relatively slow. When performing access is realized through a global mechanism. The control
dangerous operations in the virtual client, the host computer and management of the system is based on a single system
can close the connection with the virtual client and close the strategy, and the global benefits are considered. Because of
virtual client at any time, thus minimizing the possibility of the encapsulation caused by virtualization, snapshots
hardware damage or system paralysis caused during the become a feature of virtual machine disk mirroring. Using
experiment. snapshots, virtual machine disk mirroring can be restored
from the wrong state to the correct state at any moment.
The relationship between distributed computer system DNOS is a set of software packages copied on each node
and computer network is easy to be blurred. As far as we
machine of distributed system, and runs together on each
know now, there are similarities but essential differences independent node machine, providing a virtual machine for
between them. Before establishing a system design method,
distributed system users by providing a set of facilities.
the design goal of the system should be defined first. The
network infrastructure is shown in Figure 1. Global internal mode consists of three parts, namely,
slice mode, allocation mode and mapping mode. The local
conceptual model represents a local database with the data
description language of the local data model. The global
structure is shown in Figure 2.

Figure 1 Network infrastructure

In a computer network, each user uses his own machine

in a distributed computer system, and processes are
dynamically scheduled among computers in the system. In
the computer network, users know where their files are in
the distributed computer system, and the file placement is Figure 2 Global view structure
managed by the operating system. When the virtual client is
completely damaged in the experiment, the backup file can The remote deployment of computer operating systems
be copied. Rebuild the virtual client quickly. Its based on virtual machines is actually a continuum of
reconstruction speed and flexibility are faster and better than complex links [13]. First, install the remote installation
that of restoring the operating system with cloning software. service on the server. Run the "Add/Remove Windows
Through virtual machine software, one or more virtual Components" wizard in "Add or Remove Programs", and
computers can be simulated on one physical computer. install the "Remote Installation Service" in the pop-up
These virtual computers have their own hardware devices "Windows Components Wizard" dialog box. After the
such as CPU, memory, hard disk and network card. The installation is complete, restart the computer. Second, run
meaning of transparency includes file storage and on which the "risetup" command, open the "Remote Installation
node the program is executed, which is invisible to the user, Service Installer Wizard", click "Next", specify the remote
that is, transparent. However, it is very difficult to construct installation folder location as D:\Remotelnstall, and specify
a completely transparent distributed system, which is mainly the interaction between the RIS server and the client
due to the extra overhead and delay caused by the dispersion computer in the initial settings Methods. Once again, run the
of physical resources and the tightness of logic root [12]. "dsa" command for the necessary settings on the server side,
The strong reusability of virtual machines makes it easy to open the "Active Directory Users and Computers", expand
maintain the experimental environment. After installing a the rjxy.cn domain, create a new organizational unit named
virtual client on a computer, it can be copied to other hard rjxy, and create two sub-OUs in rjxy, users and computers,
disk partitions or other computers and used directly without to accommodate User accounts and computer accounts in
modification. rjxyOU. Finally, create a virtual remote installation startup
floppy disk to run Virtual PC Console, click "File" "Virtual
III. DISTRIBUTED OPERATING SYSTEM FOR MULTIPLE Disk Wizard Finally, use Virtual PC Console to create a
VIRTUAL MACHINES new virtual machine to simulate the client computer.
We should clearly realize that the realization of this Classifying each resource entity of the system,
automatic installation actually depends on a series of abstracting the data of type resource management and
powerful support technologies including communication abstracting the control of system service functions, so that
technology between client and remote deployment server, the whole system becomes a whole set of some functions,

95

Authorized licensed use limited to: Miami University Libraries. Downloaded on June 14,2021 at 20:14:37 UTC from IEEE Xplore. Restrictions apply.
and each function completes a certain and predefined known in the system scope, while the other parts of the task
service as a whole. In the system design based on task are local to the task, and are created and generated by the
model, the interaction between tasks is accomplished management process with the continuous operation of the
through information transfer primitives. When a task whole system. The object is generated with the received
receives a service request primitive, its management process creation and deletion primitives, and the processing
searches for the required object according to the information subprocess is randomly created according to the service
given in the request and creates a corresponding processing request primitive received by the task, and is terminated by
subprocess on the required object to respond to the service the management process after executing the processing
request. If the system needs to expand its functions or request.
develop other application systems, it is realized by calling
In the network operating system, resources are managed
the existing system functions as a whole. After these two
steps of abstraction, the design of the whole system is locally, used locally, and accessed through local mechanism,
without any single system strategy. Whether using the flat
structured, modular and functional-oriented.
image format or the wizard image format for remote
In order to obtain the optimal rate of the startup process, installation, it actually needs to install files directly to the
the total throughput of all storage nodes needs to be operating system. Some of the core of DNOS are concrete
calculated. Figure 3 shows the total read rate when multiple solutions to these problems, but it is aimed at resources, and
virtual machines are started. Figure 4 corresponds to the the parts that need to be divided and abstracted are resources.
total write rate when multiple virtual machines are started. The implementation method of division and abstraction
adopts task model. In a process-oriented system, the
operating system is a collection of processes, while in an
object-oriented system, the operating system is a collection
of objects. In the process-oriented system, the
synchronization and control of user process and system state
is realized through message transmission, while in the
object-oriented system, this synchronization and control is
accomplished through the management and distribution of
capabilities. Virtual desktop system is used in enterprises to
provide a unified working interface for employees. Most
users will start their virtual machines within the time period
after work starts. In the enterprise environment, hundreds of
virtual machines are started at the same time, which will
consume the throughput of most single storage services.
IV. CONCLUSION
Figure 3 The total read rate when multiple virtual machines are started With the popularization of desktop virtualization
technology, many enterprises began to deploy desktop
virtualization systems. Intelligent computer, neural
computer and molecular computer in the future must be a
highly parallel multi-computer system. Distributed computer
system has made great progress in recent years, and has
been paid more and more attention by the scientific and
technological circles, and its application has attracted more
and more attention. For virtual machines with certain
simulation function, whether they can be widely used is
particularly important and critical. In fact, its deployment is
a thorny problem. In this kind of measures, the
corresponding countermeasures are put forward and
implemented, but they also show a complex feature. As a
result, we need the vast number of computer practitioners to
carry out the actual remote deployment of virtual machines.
Using virtual machine software to assist the training and
teaching of computer operating system, the teaching
Figure 4 The total write rate when multiple virtual machines are started demonstration and student experiment can be carried out
smoothly. As a virtual client is a file created by VMware
A user's service request may need to schedule a series of software on a physical computer, any operation on the
tasks to complete together, which together provide a virtual virtual machine will not damage the hardware devices.
machine system for the user. The interaction of objects Therefore, experiments that cannot be done in traditional
between tasks is carried out indirectly through the teaching can be completed in virtual machines.
management process in the task. After the data abstraction
of the first step object model, the system physical resource REFERENCES
entities form a data abstract entity set. An entity represents a [1] Huo Yanyan, Guan Yong, Li Xiaojuan, et al. Formal verification of
kind of resource. Now we divide and classify these entity task scheduling of distributed real-time operating system based on
sets again to form some complete but mutually exclusive PRISM. Small Microcomputer System, vol. 36, no. 9, pp. 2125-2129,
entity sets. These entity sets together with their management 2015.
process form a task. The management process in the task is [2] Lantz B, O'Connor B. A Mininet-based Virtual Testbed for

96

Authorized licensed use limited to: Miami University Libraries. Downloaded on June 14,2021 at 20:14:37 UTC from IEEE Xplore. Restrictions apply.
 Distributed SDN Development. ACM SIGCOMM Computer Storage System Optimization Technology Based on User-Level
Communication Review, vol. 45, no. 5, pp. 365-366, 2015. Fusion I/O. Computer Research and Development, vol. 057, no. 003,
[3] Hou Yufeng, Ni Ming. Research and Implementation of Satellite pp. 649-659, 2020.
Group Distributed Technology. Computer System Applications, vol. [9] Lu Wei, Zhang Longmei. Research on Hybrid Task Scheduling
26, no. 12, pp. 233-239, 2017. Technology and Strategy of Embedded Operating System. Computer
[4] Yuan Yue. A hierarchical cloud operating system performance Engineering and Applications, vol. 51, no. 15, pp. 6-11, 2015.
diagnosis method. Computer Science, vol. 46, no. 3, pp. 327-332, [10] Bal H, Epema D, De Laat C, et al. A Medium-Scale Distributed
2019. System for Computer Science Research: Infrastructure for the Long
[5] Guo Zhenghong, Zhao Bingchen. Internet of Things Architecture Term. Computer, vol. 49, no. 5, pp. 54-63, 2016.
Model and Application Based on Web Services. Computer [11] Tamura Y, Doan T T, Chiba T, et al. A real-time operating system
Applications and Software, vol. 32, no. 5, pp. 123-126, 2015. supporting distributed shared memory for embedded control systems.
[6] Kawamura Y, Tahara J, Kato T, et al. Control of μ-Autonomous Cluster Computing, vol. 13, no. 10, pp. 1-10, 2017.
Surface Vehicle Using Sliding Mode Control. Sensors and materials, [12] Tamura Y, Doan T T, Chiba T, et al. A real-time operating system
vol. 31, no. 12, pp. 4231-4245, 2019. supporting distributed shared memory for embedded control systems.
[7] Silva A M B D, Guimaraes G C, Chagas M L R, et al. Performance Cluster Computing, vol. 22, no. 1, pp. 989-998, 2019.
Analysis of Distributed Synchronous Generators with Controllers [13] Liang Mincang, Yin Yong, Ding Tianming, et al. Design and
Equipped with Switchable Operating Modes. IEEE Latin America Application of Distributed Deep-Submersible Retraction Operation
Transactions, vol. 14, no. 5, pp. 2280-2290, 2016. Simulation System. Computer Simulation, vol. 37, no. 5, pp. 306-310,
[8] An Zhongqi, Zhang Yunyao, Xing Jing, et al. Research on Key-Value 407, 2020.

97

Authorized licensed use limited to: Miami University Libraries. Downloaded on June 14,2021 at 20:14:37 UTC from IEEE Xplore. Restrictions apply.