+

Characterization of Distributed
Systems
The components of a DS are located at networked computers
Except where otherwise indicated, this work is © 2016 by José and communicate and coordinate by passing messages
María Foces Morán and its contents is mainly based on the
textbook by Coulouris, Dollimore, Kindberg and Blair. Distributed
Systems: Concepts and Design, Edition 5, © Addison-Wesley 2012
15/9/21 V4.5
+ 2

Chapter objectives

1. Introduction

2. Examples of distributed systems

3. Trends in distributed systems

4. Focus on resource sharing

5. Challenges

6. Case study: The World Wide Web

+ 3

Study of this chapter

1. Leverage the value of this lecture:

n Attempt solving the questions about Networking included at the beginning of Questionnaire SG-
DS-1.0.pdf (Questionnaire no.1 of DS; paloalto.unileon.es/ds/SG/SG-DS-1.0.pdf).
n Take class notes

2. Have the textbook by Kindberg et al. at hand

n The physical book is in our Library
n A pdf versión is available

3. After the lecture, solve all the questions included in the Questionnaire mentioned
above
n If you need assistance
n After that, bring your questions to the next B1 session
n Attend one of the Tutor Teaching sessions
+ 4

What is a Distributed System?

n A Distributed System is made up of a set of networked computers (physically

separated) that coordinate their actions by only interchanging messages and which
objective consists of sharing resources

n Processes running at different hosts are concurrent (Actually parallel)

n Hosts coordinate by exchanging messages, only

n Accuracy of clock synchronization is limited, no global clock

n Independent failures: Other nodes may still run after one fails

n Resource sharing is the main motivation

+ 5

Examples of Distributed Systems

Finance and commerce eCommerce e.g. Amazon and eBay, PayPal, online
banking and trading
The information society Web information and search engines, ebooks, Wikipedia;
social networking: Facebook and MySpace.
Creative industries and online gaming, music and film in the home, user-
entertainment generated content, e.g. YouTube, Flickr
Healthcare health informatics, on online patient records,
monitoring patients
Education e-learning, virtual learning environments; distance
learning
Transport and logistics GPS in route finding systems, map services: Google
Maps, Google Earth
Science The Grid as an enabling technology for collaboration
between scientists
Environmental management sensor technology to monitor earthquakes, floods or
tsunamis
From Coulouris, Dollimore, Kindberg and Blair. Distributed Systems: Concepts and Design, Edition 5, © Addison-Wesley
2012
+ 6

Examples: Google Web search

n Index the contents of the entire web (63 billion pages)

n A major challenge, one of the most complex DS in history
n Highlights
n Large number of nodes
n Distributed file system
n Distributed storage system
n Distributed file locking
n Parallel programming allowed
+ 7

Examples: MMOGs

n Massively Multiplayer Online Games

n Over 50000 simultaneous players

n Fast response times

n Real time propagation of events to many players

n Solutions proposed:
n 1. Client-server architecture, a single copy of the state of the world
n Server: A powerful cluster
n Load is partitioned
n 2. Geographycally distributed architecture (Physical. Proximity)
n 3. Radically new architectures based on peer-to-peer computing (decentralized)
+ 8

Examples: Financial trading

n Cutting edge distributed systems used in this industry

n Essential: Reliable and and timely delivery of events

n Architecture: Distributed event-based systems
n FIX (Financial Information eXchange protocol)
n Support heterogeneity
n Adapter systems
n Real time processing to detect patterns of trading

n Manage risk

n Regulations compliance

n Fraudulent transactions

From Coulouris, Dollimore, Kindberg and Blair. Distributed Systems: Concepts and Design, Edition 5, © Addison-Wesley
2012
+ 9

Trends in DS

DS are undergoing significant changes at present

n Pervasive Networking and Internet

n Mobile and Ubiquitous Computing

n Distributed Multimedia Systems

n Distributed Computing as a Utility

+ 10

Pervasive Networking, Modern Internet

n Modern networking technologies

n WiFI, WiMax, LTE, MPLS

n Networking is becoming pervasive

n Affecting more and more aspects of persons lives

n Protocols
n Abstracting a myriad of technologies
n Programs can run anywhere and send messages
to any node
Network Convergence. Vinod Joseph, Srinivas Mulugu. MKP 2014

n Web ≠ Internet

n Intranets, firewalls: filter incoming and outgoing

messages
+ 11

Mobile Computing

• Users who are away from their ‘home’ network

• Can still access shared resources on their home networks

• Mobile IP:
Correspondent
host

Foreign
Network
Home Internet
Network
HoA FoA

Mobile node
+ 12

Ubiquitous Computing

• Connecting the cheap

computing devices present in
the physical environment
• Home, office and natural things
• IOT = Internet Of Things
• IOE = Internet Of Everything

• Spontaneous interoperation
• Service discovery
• Device associations
+ 13

Distributed Multimedia Systems

n Multimedia
n Discrete media
n Pictures
n Text
n Continuous media (Temporal
dimension, considerable demands)
n Audio
n Video
n IP telephony
n WebCasting
n Video Conferencing Network Convergence. Vinod Joseph, Srinivas Mulugu. MKP 2014
+ 14

Quality of Service (QoS)

n Applies to Network and Operating Systems alike

n Non-functional requirements
n Reliability
n Security
n Performance
n Response time
n Throughput
n Delay and jitter
n Loss
+ 15

Distributed computing as a utility

n Distributed resources as a commodity or a utility

n Regardless of company or brand
n Water supply
n Electricity
n Wheat

n Resources are rented, not owned

n Physical resources
n Sophisticated Data Centers
n OS virtualization:
n Virtual nodes, not physical nodes
n Flexibility From Coulouris, Dollimore, Kindberg and Blair. Distributed Systems: Concepts and Design, Edition 5, © Addison-Wesley
2012

n Software services
+ 16

Utility computing = Cloud computing

n Everything is a service

n Reduces requirements of users devices

n Implemented on cluster computers

n CLUSTER: High speed LAN
interconnection of hosts collaborating in
speeding up a large computational load
n Blade servers in a rack

From Coulouris, Dollimore, Kindberg and Blair. Distributed Systems: Concepts and Design, Edition 5, © Addison-Wesley
2012
+ 17

Focus: Resource Sharing

n When Distributed Systems are programmed in an Object-

Oriented language (Java; C++)
n Resources are encapsulated as objects
n Client
objects
n Server objects
+ 18

Challenges:
Heterogeneity
• Networks • What about data types?
• Before data can be exchanged
• Hardware there must be
an agreement about their physical
• OS representation: Marshalling

• Languages • Middleware
• CORBA is a specification of which
• Implementations many implementations exist
• Masked by the abstractions of • JAVA RMI is a part of the Java
Communication Protocols platform
+ 19

Challenges: Openness

Refers to DS components specification and documentation

n If key interfaces are public, then:

n New implementation of DS is possible
n RFCs

n Open distributed systems are extensible

n What is then an Open Source System? Is an OSS open?

+ 20

Challenges: Security and scalability

n Of paramount importance today

n Confidentiality
n Integrity
n Availability
n Authentication

n Scalability
n System performance will be O(n) if more resources are added
n n users à O(n) servers then, system is scalable
+ 21

Failures in DS

• Detecting failures • Redundancy

• CRC, checksum • Redundant components
• Two routes to any
• Tolerating failures
router
• Recovery • DNS
• Databases

• High availability
+ 22

Concurrency
n Several clients will attempt to access
a resource simultaneously on a server
n If server serves one request at a time:
n Limits throughput
n But, will guarantee data consistency

n Ifserver can serve several requests

concurrently
n Improves thorughput, but
n Concurrent operations (By several threads)
on one object may interact among themselves
and produce inconsistent results
n Non-deterministic, in general
+ 23

Distributed System transparencies

The separation of components of a DS remains hidden to the user and the programmer

o Access transparency

Enables local and remote resources to be accessed using

identical operations

o Location transparency

Enables resources to be accessed without knowledge of

their physical or network location (for example, which
building or IP address)
+ 24

Distributed-System Transparencies
The separation of components of a DS remains hidden to the user and the programmer

o Concurrency
Enables several processes to operate concurrently using shared
resources without interference between them

o Replication
Enables multiple instances of resources to be used to increase
reliability and performance without knowledge of the replicas by users
or application programmers
+ 25

Distributed-System Transparencies
The separation of components of a DS remains hidden to the user and the programmer

o Failure
Enables the concealment of faults, allowing users and application
programs to complete their tasks despite the failure of hardware or
software components.

o Mobility
Allows the movement of resources and clients within a system without
affecting the operation of users or programs
+ 26

Distributed-System Transparencies
The separation of components of a DS remains hidden to the user and the programmer

o Performance
Allows the system to be reconfigured to improve performance as
loads vary.

o Scaling
Allows the system and applications to expand in scale without change
to the system structure or the application algorithms.
 +

15/9/21 27