Scribd
Upload
66 views

02 SystemModels Architecture 2

The document discusses various models for distributed system architectures including interaction models, failure models, and security models. It then covers challenges in distributed systems like varying usage patterns and threats. Different architectural models are presented including client-server, multi-tiered architectures, and peer-to-peer systems.

Uploaded by

Kool Prashant
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
66 views

02 SystemModels Architecture 2

The document discusses various models for distributed system architectures including interaction models, failure models, and security models. It then covers challenges in distributed systems like varying usage patterns and threats. Different architectural models are presented including client-server, multi-tiered architectures, and peer-to-peer systems.

Uploaded by

Kool Prashant
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

System Models –

Distributed Architecture

Adapted from From


DistributedSystems: Concepts and Design, Coulouris,
Dollimore and Kindberg
Edition 4, © Addison-Wesley 2005

Distributed System Models


 Architectural Models – Placement of parts in a
distributed system and the relationship between them.

 Fundamental Models – Description of properties that


are present in all distributed architectures.
 Interaction Models – Issues dealing with the interaction of
process such as performance and timing of events.
 Failure Models – Specification of faults that can be exhibited by
processes and communication channels.
 Security Models – Threats to processes and communication
channels

1
Distributed System Challenges

 Varying modes of use:


 Load, connectivity, conflicting requirements
 Wide range of system environments:
 Heterogeneous HW, OS, networks, scale
 Internal challenges:
 Non-synchronized clocks, data updates, many modes
of HW & SW failure modes
 External threats:
 Attacks on data integrity & secrecy, denial of service

Architecture Models

 Distributed Architectures:
 Placement of components across a network for
the useful distribution of data and workload
 Define functional roles of components and
patterns of communication between them
 Driven by non-functional requirements:
 Performance, reliability, security, cost, etc.

2
Software and hardware service layers in distributed systems

Applications, services

Middleware

Operating system
Platform

Computer and network hardware

Software and hardware service layers in


distributed systems

 Platform :
 Box dependent HW & SW layers that provide services
to OS layers above
 Middleware:
 Layer of software that masks heterogeneity and
provides a convenient programming model for
application programmers.
 Java RMI, CORBA, web services,DCOM(.NET Remoting)
 Services – transactions, persistence, naming, etc.

3
Limitations of Middleware

 Simplifies much distributed development,


but still requires support at application level
 “end-to-end argument”
 Saltzer, Reed & Clark (1984)
 All communication activities cannot be
completely abstracted away
 Affects design decisions

Client–Server Model

Client invocation Server


invocation

result result
Server

Client
Key:
Process: Computer:

• Historically the most important and most widely used


• Servers may in turn be clients of other servers
• Simple approach to sharing data and resources, but scales poorly
(bounded by capacity of server host & bandwidth of network)

4
Variation -a service provided by multiple servers

Service

Server

Client

Server

Client
Server

• Partition set of objects, or replicate copies on several hosts


• Web as an example

Variation - Web proxy server

Client Web
server
Proxy
server

Client Web
server

• Supports concept of a cache – store of recently used data objects


• Web browser cache
• Web proxy servers – shared cache of resources for multiple clients
• Increases availability and performance by reducing load on network
and web servers.

5
Variation - Thin clients and compute servers

Compute server
Network computer or PC

Thin network Application


Client Process

• Thin client – supports a window based interface only


• Compute server – cluster or multiprocessor computer running the application
• Classic example - X-11 Window System
• Highly interactive graphical applications tax the network and compute server

A distributed application based on peer


processes

Application Application

Coordination Coordination
code code

Application

Coordination
code

• Processes interact cooperatively as peers – no client/server distinction


• Data and resources are collectively stored and managed
• Addresses client/server boundary of computing and communication load
• Objects replicated to distribute load and increase availability
• Classic example - Napster

6
Design Requirements for Distributed Architectures

 Performance Issues
 Response Time – time system takes to process a
request.
 Responsiveness – how quickly system acknowledges a
request (as opposed to processing it)
 Throughput – rate at which work gets done
 Latency – time needed to accomplish remote requests
 Note that “performance” can be viewed from
conflicting perspectives (throughput vs response time)

Guidelines for Designing for Performance


“Make it work, make it right, make it fast” – Kent Beck

 Start with a “good enough” design (make it work)

 Make the system meet functional requirements (make it right)

 Take measurements to accurately access performance – refactor as


needed (make it fast)

 Don’t forget maintainability issues – tradeoff


readability/understandability against performance.

http://c2.com/cgi/wiki?MakeItWorkMakeItRightMakeItFast

You might also like