CSCE455/855 Distributed Operating Systems: Dr. Ying Lu Schorr Center 106

CSCE455/855 Distributed Operating Systems
Introduction
Dr. Ying Lu
ylu@cse.unl.edu
Schorr Center 106
CSCE455/855 Distributed Operating Systems
Giving credit where credit is due:

 Most of the lecture notes are from the textbook
companion website
 Some of the lecture notes are based on slides
created by Dr. Zahorjan at Univ. of Washington
and Dr. Konev at Univ. of Liverpool
 I have modified them and added new slides
What is an Operating System?
 The text:
 “an intermediary between the user of a computer
and the computer hardware”
 “manages the computer hardware”
 “an amazing aspect of operating systems is how
varied they are in accomplishing these tasks …
mainframe operating systems … personal
computer operating systems … operating systems
for handheld computers …”
What is an Operating System?
 An operating system (OS) is:

 a software layer to abstract away and manage details
of hardware resources
 a set of utilities to simplify application development
Applications
OS
Hardware
The major OS issues
 structure: how is the OS organized?
 sharing: how are resources shared across users?
 naming: how are resources named (by users or
programs)?
 security: how is the integrity of the OS and its
resources ensured?
 protection: how is one user/program protected from
another?
 performance: how do we make it all go fast?
More OS issues…
 concurrency: how are parallel activities (computation

and I/O) created and controlled?
 scale: what happens as demands or resources increase?
 persistence: how do you make data last longer than
program executions?
 reliability: what happens if something goes wrong
(either with hardware or with a program)?
 extensibility: can we add new features?
 flexibility: are we in the way of new apps?
 communication: how do programs exchange
information?
Course Aims
 Provide an understanding of the technical issues

involved in the design of modern distributed
systems.
 Appreciation of the main principles underlying

distributed systems: processes, communication,
naming, synchronization, consistency, fault
tolerance, and security.
Definition of a Distributed System (1)
 A collection of independent computers that

appears to its users as a single coherent
system.
 Two aspects of this definition:

1. Hardware: the machines are autonomous.
2. Software: the users think they are dealing a single system
(which is achieved by software).
Examples of Distributed Systems
 Computer world:
 University computer network
 GRID (distributed computing facilities)
 Ordinary life:
 WWW, P2P systems (such as Napster)
 Banks (Cash machines)
 Ticket reservation
Characteristics of a Distributed
System (I)
 Differences between the various computers
and the way in which they communicate are
hidden from users.
 Users and applications can interact with a
distributed system in a consistent and uniform
way, regardless of where and when
interaction takes place.
Characteristics of a Distributed
System (II)
 As a direct consequence of having
independent computers and hiding their
differences, a distributed system should be
relatively easy to expand or scale.
 A distributed system will normally be
continuously available, although perhaps
certain parts may be temporarily out of order.
Definition of a Distributed System (2)
Figure 1-1. A distributed system organized as middleware. The

middleware layer extends over multiple machines, and offers
each application the same interface.
Goals of Distributed Systems
 Easily connect users/resources
 Exhibit transparency
 Support openness
 Be scalable
Connecting Users and Resources
 Typical resources
 Printers, computers, computing power, data
 Why sharing
 Economics
 Collaboration, information exchange (groupware)
 Problems with sharing
 Security
 Unwanted collaboration
Transparency in a Distributed System
 Transparent distributed system:

 Looks to its users as if it were only a single
computer system
Transparency in a Distributed System
Figure 1-2. Different forms of transparency in a

distributed system (ISO, 1995).
Degree of transparency
 Distribution transparency is generally preferable, but

is not always a good idea:
 It is undesirable to hide the location of a shared
printer from its users.
 There is a need for trade-off between a high degree of
transparency and the performance of a system.
 It is impossible to hide the fact that Mother Nature
will not allow it to send a message from one process
in San Francisco to the other in Amsterdam in less
than approximately 35 ms.
Openness
 Open systems
 Offer services according to standard rules that
describe the syntax and semantics of these services
 Enjoy neutral and complete specifications
 Network protocols
 Advantage of being open:
 Interoperability: open systems can work together
 Portability: ability to transform an application
from one software or hardware platform to another
Scalability
 Along three different dimensions
 Size (the number of users and/or processes)
 Geographical (maximum distance between
participants)
 Administrative (number of administrative
domains)
Centralized Solutions: Obstacles for
Achieving Size Scalability
Concept Example
Centralized services A single server for all users
Centralized data A single on-line telephone book

Centralized Doing routing based on complete
algorithms information
Examples of scalability limitations.
How scalability is accomplished for

routing in the Internet?
Characteristics of Decentralized
Algorithms
 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
 Three is no implicit assumption that a global clock exists

Difficulties for Achieving
Geographical Scalability
 Synchronous communication suitable for LAN is
unsuitable for WAN due to different communication
latency
 Communication in WAN is inherently unreliable and

virtually always point-to-point, whereas LAN
communication is generally highly reliable and based
on broadcasting (e.g., location service)
 Centralized solutions lead to a waste of network

resources and degrade system performance
Difficulties for Achieving
Administrative Scalability
 Administrative scalability is hindered by conflicting

resource usage, management, and security policies in
multiple, independent administrative domains
Pitfalls when Developing
Distributed Systems
 False assumptions made by first time developer:
• The network is reliable.
• The network is secure.
• The network is homogeneous.
• The topology does not change.
• Latency is zero.
• Bandwidth is infinite.
• Transport cost is zero.
• There is one administrator.
Scaling Techniques
 Three techniques:
 Hiding communication latencies
 Try to avoid waiting for responses to remote service
requests as much as possible
 Distribution
 Replication
Scaling Techniques (1)
1.4
How to hide the communication latencies?

How to avoid waiting for the responses to
remote service request as much as possible?
1.4
The difference between letting a server or a client check

forms as they are being filled
Shipping code approach widely supported by the Web in
the form of Java applets and Javascript
1.5
An example of dividing the DNS name space into zones.

e.g. resolve name flits.cs.vu.nl
 World Wide Web
 An enormous document-based information
system
 How scalability is achieve?

 distribution of documents
 Replication
 Increases availability
 Balances the load
 Reduces communication latency
 But causes consistency problems
 Many existing consistency maintenance solutions are
inherently non-scalable
 Caching (client-driven)
Why Middleware is an Important
Component of Distributed Systems
 Distributed system
 A collection of independent computers that
appears to its users as a single coherent system.
 Two aspects of this definition:
1. Hardware: the machines are autonomous.
2. Software: the users think they are dealing a single system
(which is achieved by software).
Middleware was invented to meet practical

needs for building distributed systems
Cluster Computing Systems
 Figure 1-6. An example of a cluster computing

system.
Grid Computing Systems
 Figure 1-7. A layered architecture for grid computing systems.

Transaction Processing Systems (1)
 Figure 1-8. Example primitives for

transactions.
 Characteristic properties of transactions (ACID):
• Atomic: To the outside world, the transaction happens
indivisibly.
• Consistent: The transaction does not violate system
invariants.
• Isolated: Concurrent transactions do not interfere with
each other.
• Durable: Once a transaction commits, the changes are
permanent.
 Figure 1-9. A nested transaction.

 Figure 1-10. The role of a TP monitor in

distributed systems.
Enterprise Application Integration
 Figure 1-11. Middleware as a communication facilitator in enterprise

application integration.
Distributed Pervasive Systems
 Home Systems
 Location/context aware
 Electronic Health Care Systems
 Security
 Sensor Networks
 Data storage/aggregation/query
Sensor Networks (2)
 Figure 1-13. Organizing a sensor network database, while

storing and processing data (a) only at the operator’s site or …
Sensor Networks (3)
 Figure 1-13. Organizing a sensor network database, while

storing and processing data … or (b) only at the sensors.

CSCE455/855 Distributed Operating Systems: Dr. Ying Lu Schorr Center 106

Uploaded by

Copyright:

Available Formats

CSCE455/855 Distributed Operating Systems: Dr. Ying Lu Schorr Center 106

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CSCE455/855 Distributed Operating Systems: Dr. Ying Lu Schorr Center 106

Uploaded by

Copyright:

Available Formats

CSCE455/855 Distributed Operating Systems

Giving credit where credit is due:

 An operating system (OS) is:

 concurrency: how are parallel activities (computation

 Provide an understanding of the technical issues

 Appreciation of the main principles underlying

 A collection of independent computers that

 Two aspects of this definition:

Figure 1-1. A distributed system organized as middleware. The

 Transparent distributed system:

Figure 1-2. Different forms of transparency in a

 Distribution transparency is generally preferable, but

Centralized services A single server for all users

Centralized data A single on-line telephone book

Examples of scalability limitations.

How scalability is accomplished for

 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

 Three is no implicit assumption that a global clock exists

 Communication in WAN is inherently unreliable and

 Centralized solutions lead to a waste of network

 Administrative scalability is hindered by conflicting

How to hide the communication latencies?

The difference between letting a server or a client check

An example of dividing the DNS name space into zones.

 How scalability is achieve?

Middleware was invented to meet practical

 Figure 1-6. An example of a cluster computing

 Figure 1-7. A layered architecture for grid computing systems.

 Figure 1-8. Example primitives for

 Figure 1-9. A nested transaction.

 Figure 1-10. The role of a TP monitor in

 Figure 1-11. Middleware as a communication facilitator in enterprise

 Figure 1-13. Organizing a sensor network database, while

 Figure 1-13. Organizing a sensor network database, while

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.