1

Introduction

Chapter 1

2

The Textbook Andrew S. Tanenbaum

& Maarten van Steen, Distributed Systems: Principles and Paradigms, Prentice Hall, 2002.

全華科技圖書 , (03)401-5467, M: 0952296068

3

Grade Counting Rule Midterm Exam 30% Final Exam 30% Roll call 10% (base grade 80, 5 times during

the term) Homework or Report 30% TA: Semmer 孫瑞祥 3282,

4

Definition of a Distributed System (1)

A distributed system is:

A collection of independent computers that appears to its

users as a single coherent system.

5

Definition of a Distributed System (2)

A distributed system organized as middleware.Note that the middleware layer extends over multiple machines.

1.1

6

The Goals of DS Connecting Users and Resources: To make it easy

for users to access, remote resources, and to share them with other users in a controlled way.

Transparency: To hide the act that its processes and resources are physically distributed across multiple computers. Definition: A distributed system that is able to present

itself to users and applications as if it were only a single computer system is said to be transparent.

Openness: A system that offers services according to standard rules that describe the syntax and semantics of those services.

Scalability:

7

Transparency in a Distributed System

Different forms of transparency in a distributed system.

Transparency Description

AccessHide differences in data representation and how a resource is accessed

Location Hide where a resource is located

Migration Hide that a resource may move to another location

RelocationHide that a resource may be moved to another location while in use

ReplicationHide that a resource may be shared by several competitive users

ConcurrencyHide that a resource may be shared by several competitive users

Failure Hide the failure and recovery of a resource

PersistenceHide whether a (software) resource is in memory or on disk

8

Scalability Problems

Examples of scalability limitations.

Concept Example

Centralized services A single server for all users

Centralized data A single on-line telephone book

Centralized algorithmsDoing routing based on complete information

9

Decentralized Algorithm’s Characteristics No machine has complete information

about the system state. Machines make decisions based only on

local information. Failure of one machine does not ruin the

algorithm. There is no implicit assumption that a global

clock exists. No way to get a globally synchronized time. In LAN, they are based on synchronous

communication.

10

Scaling Techniques There are three basic techniques for DS

scaling: Hiding communication latencies

Try to avoid waiting for responses to remote service requests as much as possible.

Using asynchronous communication. Distribution

Splitting a component into smaller parts, and subsequently spreading those parts across the system.

For example, the Internet DNS Replication

Divide attention But caching and replication may lead to consistency

problem.

11

Scaling Techniques (1)

1.4

The difference between letting:

a) a server or

b) a client check forms as they are being filled

12

Scaling Techniques (2)

1.5

An example of dividing the DNS name space into zones.

13

Hardware Concepts

1.6

Different basic organizations and memories in distributed computer systems

14

Multiprocessors (1) A bus-based multiprocessor.

1.7

15

Multiprocessors (2)a) A crossbar switchb) An omega switching network

1.8

16

Homogeneous Multicomputer Systems System Area Networks (SANs)

The nodes are mounted in a big rack and are connected through a single, often high-performance interconnection network.

Two popular connection types of SAN Mesh Hypercube

17

Homogeneous Multicomputer Systems Other samples Massively Parallel Processors (MPPs)

Consisting of thousands of CPUs High-performance interconnection network Fault tolerance is required

Clusters of Workstations (COWs) A collection of standard PCs or workstations

connected through off-the-shelf communication components such as Ethernet.

18

Heterogeneous Multicomputer Systems Distributed ASIC Supercomputer (DAS)

a wide-area distributed cluster designed by the Advanced School for Computing and Imaging (ASCI).

Consisting of four clusters of multicomputers (64 nodes each), interconnected through a ATM-switched backbone.

19

Software Concepts

Tightly-coupled DOS (Distributed Operating Systems) Used for managing multiprocessors and homogeneous multicomputers.

Loosely-coupled NOS (Network Operating Systems) Used for hetergeneous multicomputer systems Distinction from traditional OS: local services are made available to

remote clients. Middleware

System Description Main Goal

DOSTightly-coupled operating system for multi-processors and homogeneous multicomputers

Hide and manage hardware resources

NOSLoosely-coupled operating system for heterogeneous multicomputers (LAN and WAN)

Offer local services to remote clients

MiddlewareAdditional layer atop of NOS implementing general-purpose services

Provide distribution transparency

20

Distributed Operating Systems Two types of DOSs

Multiprocessor operating system Multicomputer operating system

Uniprocessor Operating System Like a virtual machine to applications Kernel mode

Can access memory and registers, and execute instructions

User mode Memory and register access is restricted.

21

Uniprocessor Operating Systems

Separating applications from operating system code through

a microkernel.

1.11

22

Uniprocessor Operating Systems (cont.) Benefits to using microkernels

Flexibility A large part of the OS is executed in user mode, it is

relatively easy to replaced a module without having to recompile or re-install the entire system.

Could be placed on different machines

Disadvantages of microkernels Due to the well-entrenched status quo Have extra communication overheads (about

20% performance degradation)

23

Multiprocessor Operating Systems (1) A monitor to protect an integer against concurrent access.

monitor Counter {

private:

int count = 0;

public:

int value() { return count;}

void incr () { count = count + 1;}

void decr() { count = count – 1;}

}

24

Multiprocessor Operating Systems (2)

A monitor to protect an integer against concurrent access, but blocking a process.

monitor Counter {

private:

int count = 0;

int blocked_procs = 0;

condition unblocked;

public:

int value () { return count;}

void incr () {

if (blocked_procs == 0)

count = count + 1;

else

signal (unblocked);

}

void decr() {

if (count ==0) {

blocked_procs = blocked_procs + 1;

wait (unblocked);

blocked_procs = blocked_procs – 1;

}

else

count = count – 1;

}

}

25

Multicomputer Operating Systems (1) General structure of a multicomputer operating

system

1.14

26

Multicomputer Operating Systems (2) Alternatives for blocking and buffering in message passing.

1.15

27

Multicomputer Operating Systems (3)

Relation between blocking, buffering, and reliable communications.

Synchronization point Send bufferReliable comm. guaranteed?

Block sender until buffer not full Yes Not necessary

Block sender until message sent No Not necessary

Block sender until message received No Necessary

Block sender until message delivered No Necessary

28

Distributed Shared Memory Systems (1/3) Programming multicomputers is much harder than

programming multiprocessors Reasons: buffering, blocking, and reliable communication, etc.

Solution: Emulating shared-memory on multicomputers system Using virtual memory capability, which is referred to Distributed

Shared Memory (DSM) DSM is achieved by page-based distributed shared memory

Some problems caused by DSM Data consistency The trade-off of page size

Larger page size causes commun. cost when memory access false Smaller page size may cause low memory hitting ratio

False sharing Having data belonging to two independent processes in the same

page

29

Distributed Shared Memory Systems (2/3)

a) Pages of address space distributed among four machines

b) Situation after CPU 1 references page 10

c) Situation if page 10 is read only and replication is used

30

Distributed Shared Memory Systems (3/3) False sharing of a page between two independent processes.

31

Network Operating System (2/4) NOS

Contrast with DOS, NOS does not assume that the underlying hardware are homogeneous and that it should be managed as if it were a single system.

Different operating systems Different kernels Different hardware

More primitive than DOS Compared to DOS

Drawbacks: hard to use Login from on machine to another. Copy files from one machine to another. Changing a configuration such as password or settings.

Advantages: Easy to add or remove a machine in NOS (they are highly

independent of each other).

32

Network Operating System (2/4) General structure of a network operating system.

1-19

33

Network Operating System (3/4) Two clients and a server in a network operating system.

34

Network Operating System (4/4) Different clients may mount the servers in different places.

35

Positioning Middleware A synthetic solution between DOS and NOS:

Middleware: To place an additional layer of software between applications and the network operating system, offering a higher level of abstraction.

General structure of a distributed system as middleware.

36

Middleware Models Remote Procedure Calls (RPCs)

Hiding network communication by allowing a process to call a procedure of which an implementation is located on a remote machine.

When calling a such procedure, parameters are transparently shipped to the remote machine where the procedure is subsequently executed, after which the results are sent back to the caller.

Distributed objects Object itself located in a single machine Making its interface available on other machines

Distributed documents World wide web (WWW)

37

Middleware and Openness

In an open middleware-based distributed system, the protocols used by each middleware layer should be the same, as well as the interfaces they offer to applications.

Fig. 1-23

38

Comparison between Systems A comparison between multiprocessor operating systems,

multicomputer operating systems, network operating systems, and middleware-based distributed systems.

Item

Distributed OSNetwork OS

Middleware-based OS

Multiproc. Multicomp.

Degree of transparency Very High High Low High

Same OS on all nodes Yes Yes No No

Number of copies of OS 1 N N N

Basis for communicationShared memory

Messages Files Model specific

Resource managementGlobal, central

Global, distributed

Per node Per node

Scalability No Moderately Yes Varies

Openness Closed Closed Open Open

39

Clients and Servers General interaction between a client and a server.

40

An Example Client and Server (1) The header.h file used by the client and server.

41

An Example Client and Server (2)

A sample server.

42

An Example Client and Server (3)

A client using the server to copy a file.

1-27 b

43

Processing Level The general organization of an Internet search

engine into three different layers

1-28

44

Multitiered Architectures (1) Alternative client-server organizations (a) – (e).

1-29

45

Multitiered Architectures (2) An example of a server acting as a client.

1-30

46

Modern Architectures An example of horizontal distribution of a Web service.

1-31