Distributed Systems Architecture

1
Distributed Systems Architecture

• Presenters
• Rose Kit Turgut Tezir

2
Outline

• Introduction to Distributed System
• Characteristics of Distributed System
• Advantages of Distributed System
• Disadvantages of Distributed System
• Loosely-coupled Systems
• Tightly-coupled system
• Array Processor

3
Introduction to Distributed System

• Centralized System
• A single computer with one ore more CPUs
  processes all incoming request
• Problems with cost, reliability
• Specification and implementation are defined
  within a single system
• Distributed System
• Opposite of centralized system
• Set of separate computers that are capable of
  autonomous operation, link by a compute network.
• Enable individual computers (different location)
  to share resources in the network
• Server implementation for the same interface
  located in different servers.

4
Distributed and Centralized Computer System
5
Characteristics of Distributed System

• Resource sharing
• Resource provided by a computer which is a member
  of a distributed system can be shared by clients
  and other members of the system via a network.
• Resource Manager is a software module (based on
  set of management policies) that provides
  interfaces which enables resource to be
  manipulated by clients, Openness
• Openness
• Resource sharing services can be incorporated
  without disruption or duplication of existing
  services
• Concurrency
• All concurrent access must be synchronised to
  avoid problems such as lost update, dirty read,
  incorrect summary unrepeatable read.

6
Characteristics of Distributed System

• Scalability
• flexible to grow in size.
• Fault tolerance
• the system appropriately handles errors
• hardware redundancy, software recovery
• Transparency
• hide all unnecessary details from users
• Example Location transparency - clients do not
  need to know the location of the servers

7
Advantages of Distributed System

• Shareability
• Allows systems to use each others resources
• Expandability
• Permits new systems to be added as members of the
  overall system
• Local Autonomy
• Manage local resources
• Improved performance
• Resource replication
• Combined processing power of multiple computers
  provides much more processing power than a
  centralised system with multiple CPUs
• Improved reliability and availability
• disruption would not stop the whole system from
  providing its services as resources spread across
  multiple computers
• Potential cost reductions

8
Disadvantages of Distributed System

• Network reliance
• problems on network would disrupt activities in
  the system as a whole
• Complexities
• must be able to deal with errors from all
  computers that make up the distributed system
• Security
• private resources would be exposed to a wider
  range of potential hackers, with unauthorised
  accesses from any computers connected to the
  system
• centralised system is more secure

9
General Types of Distributed Systems

• Loosely-Coupled Systems
• Tightly-Coupled Systems
• Array Processor Systems

10
Loosely-coupled Systems

• Loosely-coupled distributed systems are
  single-user computers or workstations that are
  locally networked enabling them to access shared
  data and resources located in other server
  computers
• Most distributed systems links two types of
  computers, workstations and servers (also known
  as client/server).
• An early example of the client/server model is
  the Xerox distributed system
• Systems with different hardware and software can
  be networked together
• Advantages of loosely coupled systems over
  tightly coupled systems
• More scalable.
• Generally more reliable/fault-tolerant.
• Allows sharing of widely separated resources.

11
Loosely-coupled Systems
12
A Loosely-coupled System
Local Network




Memory
Memory
Memory
Memory
Prg A
Prg B
Prg C
Prg D
Processor
Processor
Processor
Processor
I/O
I/O
I/O
I/O
13
Tightly-coupled system

• Tightly-coupled system connects a number of
  processor into an integrated hardware system
  under the control of a single operating system.
• The operating system assigns processor and memory
  space to users tasks and allow them to run at
  the same time.
• The hardware environment includes a shared memory
  or a high-speed connection between several
  separate processor / memory systems with a
  unified virtual addressing system. This enables
  users tasks to communicate with each other and
  with the operating system.
• Advantage of tightly coupled systems over loosely
  coupled systems
• Easier to program but expensive (Limited by the
  memory bandwidth)

14
A Tightly-coupled Multiprocessor System
Shared Memory
Program A
Program B
Program C
Program D
Cache Memory
Cache Memory
Cache Memory
Cache Memory
Processor
Processor
Processor
Processor
Input / Output
15
Array Processor

• An array processor is similar to a personal
  computer with a large number of arithmetic and
  logic units linked in a regular array. They can
  be used to perform operations in parallel on
  arrays of data.
• The distinguishing characteristics of such a
  machines is that the entire array of processors
  obeys a single stream of instructions, some of
  the instructions being applied to many data items
  distributed throughout the array of processing
  unit.
• There are also other tightlycoupled distributed
  systems that are designed to make calculations
  faster by processing them in parallel.
• Advantages
• Useful for achieving high processing speed when
  processing large regular set of data
• Disadvantages
• Geographical distribution
• Many independent tasks need to executed at the
  same time.

16
Array Processor
M
M
M
M
M
M
M
M
M
P
P
P
P
P
P
P
P
P
Input / Output
Program
M Memory P Processor