2. 2
1.1 Introduction and Definition
before the mid-80s, computers were
Very expensive (hundred of thousands or even millions
of dollars)
Very slow
Not connected among themselves
after the mid-80s: two major developments
cheap and powerful microprocessor-based computers
appeared
computer networks
LANs at speeds ranging from 10 to 1000 Mbps
WANs at speed ranging from 64 Kbps to gigabits/sec
3. 3
Definition of a Distributed System
a distributed system is:
a collection of independent computers that appears to its
users as a single coherent system - computer (Tanenbaum
& Van Steen)
this definition has two aspects:
1. hardware: autonomous(independent) machines
2. software: a single system view for the users
4. 4
Why Distributed?
Resource and Data Sharing
printers, databases, multimedia servers, ...
Availability, Reliability
the loss of some instances can be hidden
Scalability, Extensibility
the system grows with demand (e.g., extra servers)
Performance
huge power (CPU, memory, ...) available
Inherent distribution, communication
organizational distribution, e-mail, video
5. 5
Problems of Distribution
Concurrency, Security
clients must not disturb each other
Privacy
e.g., when building a preference profile
unwanted communication such as spam
Partial failure
we often do not know where the error is (e.g., RPC)
Location, Migration, Replication
clients must be able to find their servers
Heterogeneity
hardware, platforms, languages, management
6. 6
1.2 Organization and Goals of a Distributed System
to support heterogeneous computers and networks and
to provide a single-system view, a distributed system is
often organized by means of a layer of software called
middleware that extends over multiple machines
a distributed system organized as middleware; note that the middleware
layer extends over multiple machines
7. 7
Goals of a distributed system: a distributed system should
easily connect users with resources (printers, computers,
storage facilities, data, files, Web pages, ...)
reasons: economics, to collaborate and exchange
information
be transparent: hide the fact that the resources and
processes are distributed across multiple computers
be open
be scalable
Transparency in a Distributed System
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
8. 8
different forms of transparency in a distributed system
Transparency Description
Access Hide differences in data representation
(file naming, ...) and how a resource
is accessed
Location Hide where a resource is physically located; where
is http://www.google.com? (naming)
Migration Hide that a resource may move to another location
Relocation Hide that a resource may be moved to another
location while in use; e.g., mobile users using their
wireless laptops
Replication Hide that a resource is replicated
Concurrency Hide that a resource may be shared by several
competitive users; a resource must be left in a
consistent state
Failure Hide the failure and recovery of a resource
Persistence Hide whether a (software) resource is in memory
or on disk
9. 9
Openness in a Distributed System
A distributed system should be open
we need well-defined interfaces
Interoperability
components of different origin can communicate
Portability
components work on different platforms
Another goal of an open distributed system is that it
should be flexible and extensible;
easy to configure the system out of different components;
easy to add new components,
easy to replace existing ones
an Open Distributed System is a system that offers
services according to standard rules that describe the
syntax and semantics of those services; e.g., protocols in
networks
10. 10
Scalability in Distributed Systems
a distributed system should be scalable
Size: adding more users and resources to the system
Geographically: users and resources may be far apart
Administratively: should be easy to manage even if it
spans many administrative organizations
11. 11
Concept Example
Centralized services
Single server for all users-mostly for security
reasons
Centralized data A single on-line telephone book
Centralized algorithms Doing routing based on complete information
examples of scalability limitations
Scalability problems: performance problems caused by
limited capacity of servers and networks
Scaling Techniques
how to solve scaling problems
For geographical scalability: the problem is mainly
performance, due to the limitations in the capacity of
servers and networks
three possible solutions: hiding communication latencies,
distribution, and replication
12. 12
a. Hide Communication Latencies
Try to avoid waiting for responses to remote service
requests
let the requester do other useful job
i.e., construct requesting applications that use only
asynchronous communication instead of synchronous
communication; when a reply arrives the application is
interrupted
Good for batch processing and parallel applications but
not for interactive applications
for interactive applications, move part of the job to the
client to reduce communication; e.g. filling a form and
checking the entries
13. 13
(a) a server checking the correctness of field entries
(b) a client doing the job
14. 14
b. Distribution
e.g., DNS - Domain Name System
divide the name space into zones
for details, see later in Chapter 4 - Naming
an example of dividing the DNS name space into zones
15. 15
c. Replication
Replicate components across a distributed system to
increase availability and for load balancing, leading to
better performance
decided by the owner of a resource
caching (a special form of replication) also reduces
communication latency; decided by the user
but, caching and replication may lead to consistency
problems (see Chapter 6 - Consistency and Replication)
16. 16
1.3 Hardware and Software Concepts
Hardware Concepts
different classification schemes exist
multiprocessors - with shared memory
multicomputers - that do not share memory
can be homogeneous or heterogeneous
18. 18
a bus-based multiprocessor
Multiprocessors - Shared Memory
the shared memory has to be coherent –
the same value written by one processor must be read by another
processor
Performance problem for bus-based organization since the
bus will be overloaded as the number of processors
increases
The solution is to add a high-speed cache memory
between the processors and the bus to hold the most
recently accessed words; may result in incoherent memory
Bus-based multiprocessors are difficult to scale even with
caches
Two possible solutions: crossbar switch and omega
network
19. 19
Crossbar switch
divide memory into modules and connect them to the
processors with a crossbar switch
at every intersection, a crosspoint switch is opened and
closed to establish connection
problem: expensive; with n CPUs and n memories, n2
switches are required
20. 20
Omega network
use switches with multiple input and output lines
drawback: high latency because of several switching
stages between the CPU and memory
21. 21
Homogeneous Multicomputer Systems
also referred to as System Area Networks (SANs)
the nodes are mounted on a big rack and connected
through a high-performance network
could be bus-based or switch-based
bus-based
shared multiaccess network such as Fast Ethernet can be
used and messages are broadcasted
Performance drops highly with more than 25-100 nodes
(contention)
23. 23
Heterogeneous Multi-Computer Systems (HMCS)
most distributed systems are built on heterogeneous
multicomputer systems
the computers could be different in processor type,
memory size, architecture, power, operating system, etc.
and
I/O bandwidth and the interconnection network may be
highly are different
the distributed system provides a software layer to hide
the heterogeneity at the hardware level; i.e., provides
transparency
25. 25
Operating systems for distributed computers can be
roughly divided into two categories:
Tightly-coupled systems:
The OS tries to maintain a single, global view of the resources
Generally, referred to as distributed OSs (DOS)
Used for multiprocessors and homogeneous multi-computers
Each component dependent on the others
Manages all the resources in a distributed system,
Provides transparency (location, migration, concurrency,
replication)
Loosely-coupled systems, referred to as network OSs
(NOS)
In which, there are a collection of computers each running their
own operating system;
However, these OS work together to make their services and
resources available to others
Used for heterogeneous multi-computers
Software Concepts…….
26. Middleware:
To enhance the services of network operating
system (NOS)
To provide distribution transparency
To actually come to a distributed system
26
Software Concepts……
27. 27
System Description Main Goal
DOS
Tightly-coupled operating system for multi-
processors and homogeneous
multicomputers
Hide and manage
hardware
resources
NOS
Loosely-coupled operating system for
heterogeneous multicomputers (LAN and
WAN)
Offer local
services to remote
clients
Middleware
Additional layer atop of NOS implementing
general-purpose services
Provide
distribution
transparency
Summary of main issues
an overview of DOSs, NOSs, and middleware
28. 28
Distributed Operating Systems
Two types of Distributed systems
Multiprocessor operating system:
Manages the resources of a multiprocessor
Multicomputer operating system:
An operating system that is developed for
homogeneous multi-computers
29. 29
Uniprocessor Operating Systems:
One large kernel that handles everything (e.g., MS-
DOS, early UNIX).
Microkernel architecture: The OS kernel is
decomposed into micro kernels,
Separating applications from operating system code
through a microkernel
30. 30
More on Multiprocessor Operating Systems
Extended from uniprocessor operating systems to support
multiple processors having access to a shared memory
Problem: Consistency
Solution: using a protection mechanism (or two
synchronization mechanisms): semaphores and monitors
Semaphore: an integer with two atomic operations down
and up
Monitor: a programming language construct consisting
of procedures and variables that can be accessed only
by the procedures of the monitor;
only a single process at a time is allowed to execute
a procedure
31. 31
Multicomputer Operating Systems
general structure of a multicomputer operating system
Processors can not share memory; instead communication
is through message passing
Each node has its own
kernel for managing local resources
separate module for handling interprocessor
communication
32. 32
Distributed Shared Memory Systems
how to emulate shared memories on distributed systems to
provide a virtual shared memory
page-based distributed shared memory (DSM) - use the
virtual memory capabilities of each individual node
pages of address space distributed among four machines
33. 33
situation if page 10 is read only and replication is used
situation after CPU 1 references page 10
read-only pages can be easily replicated
34. 34
Network Operating Systems
possibly heterogeneous underlying hardware
constructed from a collection of uniprocessor systems, each
with its own operating system and connected to each other
in a computer network
general structure of a network operating system
35. 35
Services offered by network operating systems
remote login (rlogin)
remote file copy (rcp)
shared file systems through file servers
two clients and a server in a network operating system
36. 36
Middleware
A distributed operating system is not intended to handle a
collection of independent computers but provides
transparency and ease of use
A network operating system does not provide a view of a
single coherent system but is scalable and open
Combine the scalability and openness of network operating
systems and the transparency and ease of use of distributed
operating systems
this is achieved through a middleware, another layer of
software
38. 38
Item
Distributed OS Network
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
communication
Shared
memory
Messages Files
Model
specific
Resource management
Global,
central
Global,
distributed
Per node Per node
Scalability No Moderately Yes Varies
Openness Closed Closed Open Open
a comparison between multiprocessor operating systems,
multicomputer operating systems, network operating
systems, and middleware-based distributed systems
39. Networks
Started in 1070s
Ethernet and LAN were evented
The internet continue to be the biggest examples of distributed
systems.
Distributed systems have evolved from “LAN” based to
“Internet” based.
Telecommunication networks
Real-time systems
Flight control systems
Uber and Ride
Manufacturing automation control systems
Parallel Processing
Distributed artificial intelligence
Distributed Database Systems 39
Examples of Distributed Systems