虚拟机：系统与进程的多面平台

需积分: 13 7 浏览量更新于2024-07-18 收藏 78.64MB PDF 举报

"虚拟机：系统与进程的多功能平台" 虚拟机技术是计算机科学中的一个核心概念，它通过虚拟化技术实现对整个机器的模拟，从而克服了真实机器的兼容性和硬件资源限制，极大地提高了软件的可移植性和灵活性。这种技术在当前的计算机系统设计中扮演着至关重要的角色。虚拟机提供了系统安全、灵活性、跨平台兼容性、高可靠性以及资源效率。它们在操作系统、编程语言和计算机架构等多个领域中发挥着关键作用。在过程层面，虚拟化技术支持动态程序翻译，允许程序在不同的硬件或操作系统上运行，实现了平台无关性的网络计算。例如，Java 虚拟机（JVM）使得 Java 程序能够在任何支持 JVM 的平台上执行，这就是跨平台兼容性的体现。此外，虚拟机还支持动态加载和即时编译（JIT），提高了程序执行效率。在系统层面，虚拟机允许在同一硬件平台上运行多个操作系统环境，这对于服务器尤其有用，如在一台物理服务器上运行 Windows 和 Linux 系统，这被称为虚拟化服务器。这种方式提高了硬件利用率，降低了运营成本，并且方便进行系统隔离和故障恢复。近年来，虚拟机技术的发展受到了几个主要趋势的推动。首先，互联网的普及，尤其是万维网的崛起，催生了对跨平台解决方案的巨大需求，虚拟机成为了支持分布式计算和云服务的关键技术。例如，亚马逊的 AWS 提供了基于虚拟机的弹性计算云服务，用户可以按需创建和管理虚拟机实例。其次，移动设备和物联网（IoT）的发展也促进了虚拟机技术的应用。在移动设备上，如 Android 系统就采用了 Dalvik 虚拟机（后被 ART 取代），确保了应用的兼容性和设备的性能优化。而在 IoT 场景下，轻量级的虚拟机如 TinyVM 或者 Arm Mbed OS 上的虚拟化技术，让资源受限的设备也能运行复杂的应用。最后，容器技术的兴起，如 Docker，虽然不是传统意义上的虚拟机，但同样利用了虚拟化的概念，提供了一个轻量级的隔离环境来运行应用程序。这些容器在不引入虚拟机开销的情况下，实现了快速部署和扩展。虚拟机技术的持续发展不仅在于其技术本身的完善，也在于与新兴技术的融合，如人工智能、大数据分析和区块链等。例如，某些区块链系统如 Ethereum 就利用虚拟机来执行智能合约，保证了代码的执行一致性。虚拟机已经成为现代计算领域的基石，不仅解决了多样的系统集成问题，也为新的应用场景和创新提供了无限可能。随着技术的进步，虚拟机将更加深入地融入我们的日常生活和工作中，成为构建未来计算架构不可或缺的一部分。

odern computers are among the most advanced human-engineered

structures, and they are possible only because of our ability to manage

extreme complexity. Computer systems contain many silicon chips, each with

hundreds of millions of transistors. These are interconnected and combined

with high-speed input/output (I/O) devices and networking infrastructure to

form the platforms upon which software can operate. Operating systems,

application programs and libraries, and graphics and networking software all

cooperate to provide a powerful environment for data management, education,

communication, entertainment, and many other applications.

The key to managing complexity in computer systems is their division into

levels of abstraction

separated by

well-defined interfaces.

Levels of abstraction

allow implementation details at lower levels of a design to be ignored or sim-

plified, thereby simplifying the design of components at higher levels. The

details of a hard disk, for example, that it is divided into sectors and tracks, are

abstracted by the operating system (OS) so that the disk appears to application

software as a set ofvariable-size files (see Figure 1.1). An application program-

mer can then create, write, and read files, without knowledge of the way the

hard disk is constructed and organized.

The levels of abstraction are arranged in a hierarchy, with lower levels imple-

mented in hardware and higher levels in software. In the hardware levels, all the

components are physical, have real properties, and their interfaces are defined

so that the various parts can be physically connected. In the software levels,

components are logical, with fewer restrictions based on physical characteris-

tics. In this book, we are most concerned with the abstraction levels that are at

or near the hardware/software boundary. These are the levels where software

is separated from the

machine

on which it runs.

Chapter 1--Introduction to Virtual Machines

Figure 1.1

abstraction',,

file

]

file

I p

"',, ',

/ ,,,'"

Files Are an Abstraction of a Disk.

A level of abstraction provides a simplified interface to underlying

resources.

Computer software is executed by a

machine

(a term that dates back to

the beginning of computers;

platform

is the term more in vogue today). From

the perspective of the operating system, a machine is largely composed of

hardware, including one or more processors that run a specific instruction set,

some real memory, and I/O devices. However, we do not restrict the use of

the term

machine

to just the hardware components of a computer. From the

perspective of application programs, for example, the machine is a combina-

tion of the operating system and those portions of the hardware accessible

through user-level binary instructions.

Let us now turn to the other aspect of managing complexity: the use of

well-defined interfaces.

Well-defined interfaces allow computer design tasks to

be decoupled so that teams of hardware and software designers can work more

or less independently. The instruction set is one such interface. For exam-

ple, designers at Intel and AMD develop microprocessors that implement the

Intel IA-321 instruction set, while software engineers at Microsoft develop

compilers that map high-level languages to the same instruction set. As long

as both groups satisfy the instruction set specification, compiled software

will execute correctly on a machine incorporating an IA-32 microprocessor.

The operating system interface, defined as a set of function calls, is another

important standardized interface in computer systems. As the Intel/Microsoft

example suggests, well-defined interfaces permit development of interacting

computer subsystems at different companies and at different times, some-

times years apart. Application software developers do not need to be aware

of detailed changes inside the operating system, and hardware and software

1. Sometimes referred to informally as

x86.

1.1 Computer Architecture [] 3

can be upgraded according to different schedules. Software can run on different

platforms implementing the same instruction set.

Despite their many advantages, well-defined interfaces can also be confin-

ing. Subsystems and components designed to specifications for one interface

will not work with those designed for another. There are processors with dif-

ferent instruction sets (e.g., Intel IA-32 and IBM PowerPC), and there are

different operating systems (e.g., Windows and Linux). Application programs,

when distributed as program binaries, are tied to a specific instruction set and

operating system. An operating system is tied to a computer that implements

specific memory system and I/O system interfaces. As a general principle,

diversity in instruction sets, operating systems, and application programming

languages encourages innovation and discourages stagnation. However, in

practice, diversity also leads to reduced interoperability, which becomes restric-

tive, especially in a world of networked computers, where it is advantageous to

move software as freely as data.

Beneath the hardware/software interface, hardware resource considera-

tions can also limit the flexibility of software systems. Memory and I/O abstrac-

tions, both in high-level languages and in operating systems, have removed

many hardware resource dependences; some still remain, however. Many

operating systems are developed for a specific system architecture, e.g., for

a uniprocessor or a shared-memory multiprocessor, and are designed to man-

age hardware resources directly. The implicit assumption is that the hardware

resources of a system are managed by a single operating system. This binds

all hardware resources into a single entity under a single management regime.

And this, in turn, limits the flexibility of the system, not only in terms of avail-

able application software (as discussed earlier), but also in terms of security

and failure isolation, especially when the system is shared by multiple users or

groups of users.

Virtualization

provides a way of relaxing the foregoing constraints and

increasing flexibility. When a system (or subsystem), e.g., a processor, memory,

or I/O device, is

virtualized,

its interface and all resources visible through the

interface are mapped onto the interface and resources of a real system actu-

ally implementing it. Consequently, the real system is transformed so that

it appears to be a different, virtual system or even a set of multiple virtual

systems. Formally, virtualization involves the construction of an isomorphism

that maps a virtual

guest

system to a real

host

(Popek and Goldberg 1974). This

isomorphism, illustrated in Figure 1.2, maps the guest state to the host state

(function V in Figure 1.2), and for a sequence of operations, e, that modifies

the state in the guest (the function e modifies state

to state

Sj)

there is a

corresponding sequence of operations e ~ in the host that performs an equiv-

alent modification to the host's state (changes S I to S;). Although such an

4 9 Chapter l~Introduction to Virtual Machines

e(S i)

v(s~) v(sj)

e'(S' i)

Host

Figure 1.2

Virtualization.

Formally, virtualization is the construction of an isomorphism between a guest

system and a host; e t o V ( Si ) = V o e( Si ).

isomorphism can be used to characterize abstraction as well as virtualization,

we distinguish the two: Virtualization differs from abstraction in that virtual-

ization does not necessarily hide details; the level of detail in a virtual system is

often the same as that in the underlying real system.

Consider again the example of a hard disk. In some applications, it may be

desirable to partition a single large hard disk into a number of smaller virtual

disks. The virtual disks are mapped to a real disk by implementing each of the

virtual disks as a single large file on the real disk (Figure 1.3). Virtualizing soft-

ware provides a mapping between virtual disk contents and real disk contents

(the function V in the isomorphism), using the file abstraction as an interme-

diate step. Each of the virtual disks is given the appearance of having a number

of logical tracks and sectors (although fewer than in the large disk). A write to

a virtual disk (the function e in the isomorphism) is mirrored by a file write

and a corresponding real disk write, in the host system (the function e ~ in the

isomorphism). In this example, the level of detail provided at the virtual disk

interface, i.e., sector/track addressing, remains the same as for a real disk; no

abstraction takes place.

The concept of virtualization can be applied not only to subsystems such

as disks, but to an entire machine. A

virtual machine

(VM) is implemented

by adding a layer of software to a real machine to support the desired virtual

machine's architecture. For example, virtualizing software installed on an

Apple Macintosh can provide a Windows/IA-32 virtual machine capable of

1.1 Computer Architecture 9 5

Figure 1.3

, ",'-2-2-7-- - :----:-,

I l "''" x \

/'/?'--1"1 ~,

i~ I ))l L~ J

-'.---C" ,,

"L--'." ',

I ",..

--..- ..s

I I "..,

--.--I

I I, I 9

virtualization ~

-" ~

[ ~--

\ i I I II

\ i /

I I i

\ /

\ i /

\ I I

\ i /

\ /

\ i I /I

Implementing Virtual Disks. Virtualization provides a different interface and~or resources at the

same level of abstraction.

running PC application programs. In general, a virtual machine can circum-

vent real machine compatibility constraints and hardware resource constraints

to enable a higher degree of software portability and flexibility.

A wide variety of virtual machines exist to provide an equally wide variety

of benefits. Multiple, replicated virtual machines can be implemented on a

single hardware platform to provide individuals or user groups with their own

operating system environments. The different system environments (possibly

with different operating systems) also provide isolation and enhanced security.

A large multiprocessor server can be divided into smaller virtual servers, while

retaining the ability to balance the use of hardware resources across the system.

Virtual machines can also employ emulation techniques to support cross-

platform software compatibility. For example, a platform implementing the

PowerPC instruction set can be converted into a virtual platform running the

IA-32 instruction set. Consequently, software written for one platform will

run on the other. This compatibility can be provided either at the system level

(e.g., to run a Windows OS on a Macintosh) or at the program or process level

(e.g., to run Excel on a Sun Solaris/SPARC platform). In addition to emulation,

virtual machines can provide dynamic, on-the-fly optimization of program

binaries. Finally, through emulation, virtual machines can enable new, propri-

etary instruction sets, e.g., incorporating very long instruction words (VLIWs),

while supporting programs in an existing, standard instruction set.

The virtual machine examples just described are constructed to match

the architectures of existing real machines. However, there are also virtual

machines for which there are no corresponding real machines. It has become

剩余646页未读，继续阅读

xparmenides

粉丝: 2
资源: 33

虚拟机：系统与进程的多面平台

Virtual Machines：Versatile Platforms for Systems and Processes .part1

Virtual Machines: Versatile Platforms for Systems and Processes part3

Virtual Machines：Versatile Platforms for Systems and Processes .part2

Virtual Machines: Versatile Platforms for Systems and Processes part1

Virtual Machines: Versatile Platforms for Systems and Processes part2

Virtual Machines：Versatile Platforms for Systems and Processes .part3

Vertically Coupled Microring Resonator Filter : Versatile Building Block for VLSI Filter Circuits

Bias: Versatile Information Manager-开源

event-station:versatile通用且强大的事件发射器类

ORB-SLAM: a Versatile and Accurate Monocular SLAM System

最新资源