Grid meets Economics: A Market Paradigm for Resource Management and Scheduling in P2P/Grid Computing
Distributed System Models
Most concepts are
drawn from Chapter 2
© Pearson Education
Dr. Rajkumar Buyya
Cloud Computing and Distributed Systems (CLOUDS) Laboratory
Department of Computing and Information Systems
The University of Melbourne, Australia
http://www.cloudbus.org/652
Some ideas from Chapter 1
© Pearson Education
http://www.cs.mu.oz.au/652
2
Presentation Outline
◼ Introduction
◼ Physical Models:
◼ Three Generations of DS: Early, Internet-Scale, Contemporary
◼ Architectural Models
◼ Software Layers
◼ System Architectures
◼ Client-Server
◼ Clients and a Single Server, Multiple Servers, Proxy Servers with
Caches, Peer Model
◼ Alternative Client-Sever models driven by:
◼ Mobile code, mobile agents, network computers, thin clients, mobile
devices, and spontaneous networking
◼ Design Challenges/Requirements
◼ Fundamental Models – formal description
◼ Interaction, failure, and security models.
◼ Summary
3
Introduction
◼ Distributed systems should be designed to
function correctly in ALL circumstances/scenarios.
◼ Distributed system models helps in…
◼ ..classifying and understanding different implementations
◼ ..identifying their weaknesses and their strengths
◼ ..crafting new systems outs of pre-validated building blocks
◼ We will study distributed system models from
different perspectives
◼ Structure, organization, and placement of components
◼ Interactions
◼ Fundamental properties of systems
4
Characterization
◼ The structure and the organization of systems and
the relationship among their components should
be designed with the following goals in mind:
◼ To cover the widest possible range of circumstances.
◼ To cope with possible difficulties and threats.
◼ To meet the current and possibly the future demands.
◼ Architectural models provide both:
◼ a pragmatic starting point
◼ a conceptual view
to address these challenges.
In terms of implementation models and
basic blocks
In terms of logical view of the system,
interaction flow, and components
5
Characterization: Challenges (Difficulties and
Threats)
◼ Widely varying models of use
◼ High variation of workload, partial disconnection of components,
or poor connection.
◼ Wide range of system environments
◼ Heterogeneous hardware, operating systems, network, and
performance.
◼ Internal problems
◼ Non synchronized clocks, conflicting updates, various hardware
and software failures.
◼ External threats
◼ Attacks on data integrity, secrecy, and denial of service.
6
Characterization: Dealing with Challenges
◼ Widely varying models of use
◼ The structure and the organization of systems allow for
distribution of workloads, redundant services, and high
availability.
◼ Wide range of system environments
◼ A flexible and modular structure allows for implementing
different solutions for different hardware, OS, and networks.
◼ Internal problems
◼ The relationship between components and the patterns of
interaction can resolve concurrency issues, while structure and
organization of component can support failover mechanisms.
◼ External threats
◼ Security has to be built into the infrastructure and it is
fundamental for shaping the relationship between components.
Models at a Glance
Physical, Architectural, and
Fundamental Models
8
Physical Models
◼ A representation of the underlying H/W elements
of a DS that abstracts away specific details of the
computer/networking technologies.
◼ Baseline physical model – a small set of nodes.
◼ Three Generations of DSs (Distributed Systems):
◼ Early DSs [70-80s]: LAN-based, 10-100 nodes
◼ Internet-scale DSs [early 90-2010]: Clusters, Grids, P2P (with
autonomous nodes)
◼ Contemporary DSs: dynamic nodes in Mobile Systems that offer
location-aware services, Clouds with resource pools offering
services on pay-as-you-go basis, and Internet of Things (IoT)
(seamless interaction between physical and cyber world for
smart * applications such as Smart Health and Smart Cities)
9
Architectural model
◼ An Architectural model of a distributed system
is concerned with the placement of its parts
and relationship between them. Examples:
◼ Client-Server (CS) and Peer Process models.
◼ CS can be modified by:
◼ The partitioning of data/replication at cooperative
servers
◼ The caching of data by proxy servers or clients
◼ The use of mobile code and mobile agents
◼ The requirements to add or remove mobile devices.
10
Fundamental Models
◼ Fundamental Models are concerned with a formal
description of the properties that are common in all
of the architectural models
◼ Models addressing time synchronization, message
delays, failures, security issues are addressed as:
◼ Interaction Model – deals with performance and the
difficulty of setting of time limits in a distributed system.
◼ Failure Model – specification of the faults that can be
exhibited by processes
◼ Security Model – discusses possible threats to processes
and communication channels.
11
Presentation Outline
◼ Introduction
◼ Architectural Models
◼ Software Layers
◼ System Architectures
◼ Client-Server
◼ Clients and a Single Sever, Multiple Servers, Proxy Servers with
Caches, Peer Model
◼ Alternative Client-Sever models driven by:
◼ Mobile code, mobile agents, network computers, thin clients, mobile
devices and spontaneous networking
◼ Design Challenges/Requirements
◼ Fundamental Models – formal description
◼ Interaction, Failure, and Security models.
◼ Summary
12
Architectural Models – Intro [1]
◼ The architecture of a system is its structure in terms
of separately specified components.
◼ Its goal is to meet present and likely future demands.
◼ Major concerns are making the system reliable,
manageable, adaptable, and cost-effective.
◼ Architectural Model:
◼ Simplifies and abstracts the functions of individual
components
◼ The placement of the components across a network of
computers – define patterns for the distribution of data and
workloads
◼ The interrelationship between the components – ie.,
functional roles and the patterns of communication
between them.
13
Architectural Models – Intro [2]
◼ Architectural Model – simplifies and abstracts
the functions of individual components:
◼ An initial simplification is achieved by classifying
processes as:
◼ Server processes
◼ Client processes
◼ Peer processes
◼ Cooperate and communicate in a symmetric manner to
perform a task.
client
server
peer
peer
14
Software Architecture and Layers
◼ The term software architecture referred:
◼ Originally to the structure of software as layers or modules in a single computer.
◼ More recently in terms of services offered and requested between processes in the
same or different computers.
◼ Breaking up the complexity of systems by designing them through layers and
services
◼ Layer: a group of related functional components
◼ Service: functionality provided to the next layer.
Layer 1
Layer 2
Layer N
(services offered to above layer)
…
15
Software and hardware service layers in
distributed systems
Applicati ons, services
Computer and network hardware
Pl atform
Operating system
Middleware
16
Platform
◼ The lowest hardware and software layers are often
referred to as a platform for distributed systems and
applications.
◼ These low-level layers provide services to the layers
above them, which are implemented independently
in each computer.
◼ Major Examples
◼ Intel x86/Windows
◼ Intel x86/Linux
◼ Intel x86/Solaris
◼ PowerPC/MacOS
◼ iPhone/iOS
◼ Samsung Galaxy/Android
17
Middleware
◼ A layer of software whose purpose is to mask heterogeneity present
in distributed systems and to provide a convenient programming
model to application developers.
◼ Major Examples:
◼ Sun RPC (Remote Procedure Call)
◼ OMG CORBA (Common Object Request Broker Architecture)
◼ Microsoft D-COM (Distributed Components Object Model)
◼ Sun Java RMI (Remote Method Invocation)
◼ Modern Middleware Examples:
◼ Manjrasoft Aneka– for Cloud computing
◼ IBM WebSphere
◼ Microsoft .NET
◼ Sun J2EE
◼ Google AppEngine
◼ Microsoft Azure
18
System Architecture
◼ The most evident aspect of DS design is the
division of responsibilities between system
components (applications, servers, and other
processes) and the placement of the
components on computers in the network.
◼ It has major implication for:
◼ Performance, reliability, and security of the
resulting system.
19
Client-Server Basic Model:
Clients invoke individual servers
◼ Client processes interact with individual server processes in a separate computer
in order to access data or resource. The server in turn may use services of other
servers.
◼ Example:
◼ A Web Server is often a client of file server.
◼ Browser → search engine -> crawlers → other web servers.
Server
Cl ient
Cl ient
invocati on
result
Server
invocati on
result
Process:
Key:
Computer:
20
◼ Two-tier model (classic)
◼ Three-tier (when the server, becomes a client)
◼ Multi-tier (cascade model)
Client-Server Architecture Types
(Tier arch compliments layer architecture)
client
server
client Server/client server
client Server/client
server
Server/client
server
21
Clients and Servers
◼ General interaction between a client and a server.
22
A service provided by multiple servers
◼ Services may be implemented as several server processes in separate host computers.
◼ Example: Cluster based Web servers and apps such as Google, parallel databases Oracle
Server
Server
Server
Servi ce
Cl ient
Cl ient
23
Proxy servers (replication transparency) and
caches: Web proxy server
◼ A cache is a store of recently used data.
Cl ient
Proxy
Web
server
Web
server
server
Cl ient
24
Peer Processes: A distributed application
based on peer processes
◼ All of the processes play similar roles, interacting cooperatively as peers to
perform distributed activities or computations without distinction between clients
and servers. E.g., music sharing systems Napster, Gnutella, Kaza, BitTorrent.
◼ Distributed “white board” – users on several computers to view and interactively
modify a picture between them.
App lication
App lication
App lication
Peer 1
Peer 2
Peer 3
Peers 5 … . N
Sharable
object s
App lication
Peer 4
25
P2P with a Centralized Index Server
(e.g. Napster Architecture)
peer
peer
peer
peer
peer
peer
peer
26
Variants of Client Sever Model: Mobile code
and Web applets
◼ Applets downloaded to clients give good interactive response
◼ Mobile codes such as Applets are potential security threat, so the
browser gives applets limited access to local resources (e.g. NO
access to local/user file system).
a) cl ient request results in the downloading of appl et code
Web
server
Cl ient
Web
serverApplet
Applet code
Cl ient
b) cl ient i nteracts wi th the appl et
27
Variants of Client Sever Model: Mobile Agents
◼ A running program (code and data) that travels from one
computer to another in a network carrying out an autonomous
task, usually on behalf of some other process
◼ advantages: flexibility, savings in communications cost
◼ virtual markets, software maintain on the computers within an organisation.
◼ Potential security threat to the resources in computers they visit.
The environment receiving agent should decide which of the local
resource to allow. (e.g., crawlers and web servers).
◼ Agents themselves can be vulnerable – they may not be able to
complete task if they are refused access.
◼ Example technology:
◼ Java Agent Development Framework (JADE)
28
Thin clients and compute servers
◼ Network computer: download OS and applications from the
network and run on a desktop (solve up-gradation problem) at
runtime.
◼ Thin clients: Windows-based UI on the user machine and
application execution on a remote computer. E.g, X-11 system.
Thin
Client
Application
Process
Network computer or PC
Compute server
network
29
Mobile devices and spontaneous networking
[3rd Generation Distributed System]
◼ The world is increasingly populated by small and portable
computing devices.
◼ W-LAN needs to handle constantly changing heterogeneous,
roaming devices
◼ Need to provide discovery services: (1) registration service to
enable servers to publish their services and (2) lookup service to
allow clients to discover services that meet their requirements.
30
Summary – Models and Implications
◼ The use of CS (Client-Server) has impact on the
software architecture followed:
◼ Distribution of responsibilities
◼ Synchronization mechanisms between client and server
◼ Admissible types of requests/responses
◼ Basic CS model, responsibility is statically allocated.
◼ File server is responsible for file, not for web pages.
◼ Peer Process model, responsibility is dynamically
allocated:
◼ In fully decentralized music file sharing system, search
process may be delegated to different peers at runtime.
31
Design Requirements/Challenges of Distributed
Systems
◼ Performance Issues
◼ Responsiveness
◼ Support interactive clients
◼ Use caching and replication
◼ Throughput
◼ Load balancing and timeliness
◼ Quality of Service:
◼ Reliability
◼ Security
◼ Adaptive performance.
◼ Dependability issues:
◼ Correctness, security, and fault tolerance
◼ Dependable applications continue to work in the presence of
faults in hardware, software, and networks.
32
Presentation Outline
◼ Introduction
◼ Architectural Models
◼ Software Layers
◼ System Architectures
◼ Client-Server
◼ Clients and a Single Sever, Multiple Servers, Proxy Servers with
Caches, Peer Model
◼ Alternative Client-Sever models driven by:
◼ Mobile code, mobile agents, network computers, thin clients, mobile
devices and spontaneous networking
◼ Design Challenges/Requirements
◼ Fundamental Models – formal description
◼ Interaction, Failure, and Security models.
◼ Summary
33
Fundamental Models at Glance
◼ Fundamental Models are concerned with a formal
description of the properties that are common in all of the
architectural models
◼ All architectural models are composed of processes that
communicate with each other by sending messages over
a computer networks.
◼ Models addressing time synchronization, message
delays, failures, security issues are addressed as:
◼ Interaction Model – deals with performance and the difficulty of
setting of time limits in a distributed system.
◼ Failure Model – specification of the faults that can be exhibited by
processes
◼ Security Model – discusses possible threats to processes and
communication channels.
34
Interaction Model
◼ Computation occurs within processes;
◼ The processes interact by passing messages,
resulting in:
◼ Communication (information flow)
◼ Coordination (synchronization and ordering of activities)
between processes.
◼ Two significant factors affecting interacting
processes in a distributed system are:
◼ Communication performance is often a limiting
characteristic.
◼ It is impossible to maintain a single global notion of time.
35
Interaction Model:
Performance of Communication Channel
◼ The communication channel in our model is realised in a variety
of ways in DSs. E.g., by implementation of:
◼ Streams
◼ Simple message passing over a network.
◼ Communication over a computer network has performance
characteristics:
◼ Latency:
◼ A delay between the start of a message’s transmission from one
process to the beginning of reception by another.
◼ Bandwidth:
◼ the total amount of information that can be transmitted over in a
given time. Eg. 100 Mbps (megabits per second)
◼ Communication channels using the same network, have to share the
available bandwidth.
◼ Jitter
◼ The variation in the time taken to deliver a series of messages. It is
very relevant to multimedia data.
36
Interaction Model:
Computer clocks and timing events
◼ Each computer in a DS has its own internal clock, which can be
used by local processes to obtain the value of the current time.
◼ Therefore, two processes running on different computers can
associate timestamp with their events.
◼ However, even if two processes read their clocks at the same
time, their local clocks may supply different time.
◼ This is because computer clock drifts from perfect time and their
drift rates differ from one another.
◼ Even if the clocks on all the computers in a DS are set to the
same time initially, their clocks would eventually vary quite
significantly unless corrections are applied.
◼ There are several techniques to correct time on computer clocks.
For example, computers may use radio receivers to get readings
from GPS (Global Positioning System) with an accuracy about 1
microsecond.
37
Interaction Model:
Two variants of the interaction model
◼ In a DS it is hard to set time limits on the time taken for process
execution, message delivery or clock drift.
◼ Synchronous DS – hard to achieve:
◼ The time taken to execute a step of a process has known lower
and upper bounds.
◼ Each message transmitted over a channel is received within a
known bounded time.
◼ Each process has a local clock whose drift rate from real time has
known bound.
◼ Asynchronous DS: There is NO bounds on:
◼ Process execution speeds
◼ Message transmission delays
◼ Clock drift rates.
38
Interaction Model:
Event Ordering
◼ In many DS applications we are interested in
knowing whether an event occurred before,
after, or concurrently with another event at
other processes.
◼ The execution of a system can be described in
terms of events and their ordering despite the lack
of accurate clocks.
◼ Consider a mailing list with:
users X, Y, Z, and A.
39
Real-time ordering of events
send
receive
send
receive
m
1
m
2
2
1
3
4
X
Y
Z
Physical
ti me
A
m
3
receive receive
send
receive receive receive
t1 t2 t3
receive
receive
m
2
m1
40
Inbox of User A looks like:
◼ Due to independent delivery in message delivery, message may
be delivered in different order.
◼ If messages m1, m2, m3 carry their time t1, t2, t3, then they can
be displayed to users accordingly to their time ordering.
Item From Subject
23 Z Re: Meeting
24 X Meeting
26 Y Re: Meeting
41
Failure Model
◼ In a DS, both processes and communication
channels may fail – i.e., they may depart from
what is considered to be correct or desirable
behavior.
◼ Types of failures:
◼ Omission Failure
◼ Arbitrary Failure
◼ Timing Failure
42
Processes and channels
◼ Communication channel produces an omission failure if it
does not transport a message from “p”s outgoing
message buffer to “q”’s incoming message buffer. This is
known as “dropping messages” and is generally caused
by a lack of buffer space at the receiver or at gateway or
by a network transmission error.
process p process q
Communi cat ion channel
send
Outgoing message buffer Incoming message buffer
receivem
43
Omission and arbitrary failures
Class of failure Affects Description
Fail-stop Process Process halts and remains halted. Other processes may
detect this state.
Crash Process Process halts and remains halted. Other processes may
not be able to detect this state.
Omission Channel A message inserted in an outgoing message buffer never
arrives at the other end’s incoming message buffer.
Send-omission Process A process completes a send, but the message is not
put in its outgoing message buffer.
Receive-omission Process A message is put in a process’s incoming message
buffer, but that process does not receive it.
Arbitrary
(Byzantine)
Process or
channel
Process/channel exhibits arbitrary behaviour: it may
send/transmit arbitrary messages at arbitrary times,
commit omissions; a process may stop or take an
incorrect step.
44
Timing failures
Class of Failure Affects Description
Clock Process Process’s local clock exceeds the bounds on its
rate of drift from real time.
Performance Process Process exceeds the bounds on the interval
between two steps.
Performance Channel A message’s transmission takes longer than the
stated bound.
45
Masking Failures
◼ It is possible to construct reliable services from
components that exhibit failures.
◼ For example, multiple servers that hold replicas of data can
continue to provide a service when one of them crashes.
◼ A knowledge of failure characteristics of a
component can enable a new service to be designed
to mask the failure of the components on which it
depends:
◼ Checksums are used to mask corrupted messages.
46
Security Model
◼ The security of a DS can be achieved by
securing the processes and the channels
used in their interactions and by protecting
the objects that they encapsulate against
unauthorized access.
47
Protecting Objects: Objects and principals
◼ Use “access rights” that define who is allowed to perform operation on a
object.
◼ The server should verify the identity of the principal (user) behind each
operation and checking that they have sufficient access rights to perform
the requested operation on the particular object, rejecting those who do
not.
Network
invocati on
result
Cl ient
Server
Principal (user) Principal (server)
ObjectAccess rights
48
The enemy
◼ To model security threats, we postulate an enemy that is capable of
sending any process or reading/copying message between a pair of
processes
◼ Threats form a potential enemy: threats to processes, threats to
communication channels, and denial of service.
Communication channel
Copy of m
Process p Process qm
The enemy
m’
49
Defeating security threats: Secure channels
◼ Encryption and authentication are use to build secure channels.
◼ Each of the processes knows the identity of the principal on
whose behalf the other process is executing and can check their
access rights before performing an operation.
Principal A
Secure channelProcess p Process q
Principal B
50
Presentation Outline
◼ Introduction
◼ Architectural Models
◼ Software Layers
◼ System Architectures
◼ Client-Server
◼ Clients and a Single Sever, Multiple Servers, Proxy Servers with
Caches, Peer Model
◼ Alternative Client-Sever models driven by:
◼ Mobile code, mobile agents, network computers, thin clients, mobile
devices and spontaneous networking
◼ Design Challenges/Requirements
◼ Fundamental Models – formal description
◼ Interaction, Failure, and Security models.
◼ Summary
51
Summary
◼ Most DSs are arranged accordingly to one of a
variety of architectural models:
◼ Client-Server
◼ Clients and a Single Sever, Multiple Servers, Proxy Servers
with Cache, Peer Model
◼ Alternative Client-Sever models driven by:
◼ Mobile code, mobile agents, network computers, thin clients,
mobile devices and spontaneous networking
◼ Fundamental Models – formal description
◼ Interaction, failure, and security models.
◼ The concepts discussed in the module play an
important role while architecting DS and apps.