Lecture slides / notes.

1

Lectures on Grid Computing

Tuğba Taşkaya-TemizelProf. K. Ahmad

January 2005

2

Grid ComputingEverywhere

Business: Sectors like financial services, industrial manufacturing, energy…

Humanitarian works

Research : Health, Aerospace, Astronomy, Finance…

Government

3

Grid Computing

The internet took 20 years to be taken seriously by business. By comparison the grid is happening far more rapidly. Tom Hawk, IBM Insight Research says the worldwide market for grid technology and services is doubling every year and will reach $5 billion by 2008. Grid computing is just one of the technologies the UK government says, in its latest report, should receive more support and funding. (December 17,2003)

4

Grid Computing

"We really do believe that grid computing is real," CEO of Hewlett-Packard Carly Fiorina said. "It is driving the R&D in our industry. For the first time our energy is focused on something else than building a killer app or a hot box. We are more focused on making system that combines the best of IT and business. Imagine what is possible." (September 11, 2003)"The Grid will be the major new direction for IT," said Geoff Brown, technical director for ATS Core Technologies at Oracle. (October 28, 2002)

5

DEFINITIONS: Grid?

ELECTRICITY GRID:• A network of high-voltage transmission lines and connections that supply electricity from a number of generating stations to various distribution centres in a country or a region, so that no consumer is dependent on a single station.

UTILITY GRID: • (Term) used of any network that serves a similar purpose for other services.

www.oed.com

6

DEFINITIONS: Grid?

GRID: The Grid is envisaged to be

‘the computing and data management infrastructure that will provide the electronic underpinning for a global society in business, government, science and entertainment’

Berman, Fox and Hey (2003:9)

7

DEFINITIONS: Grid?

GRID: A virtual information

processing environment where the user has the ‘illusion’ of a seamless single-source computing power which is actually distributed.

8

Why should you care?

Ian Foster explains why we should care Grids in three points:

Vision

Reality

Future

9


Grid is a disruptive technology [Vision] It ushers in a virtualized, collaborative,

distributed world.

Two interrelated opportunities1) Enhance economy, flexibility, access by

virtualizing computing resources2) Deliver entirely new capabilities by

integrating distributed resources

Vision

Reality

Future

Vision

Reality

Future

Source: Ian Foster’ s presentation on “The Grid” , COMDEX 2003, Las Vegas, Nevada USA, November 18, 2003

10

Why should you care? Virtualization

Vision

Reality

Future

Vision

Reality

Future

Application Virtualization

• Automatically connect applications to services• Dynamic & intelligent provisioning

Infrastructure Virtualization

• Dynamic & intelligent provisioning• Automatic failover

Source: The Grid: Blueprint for a New Computing Infrastructure (2nd Edition), 2004

11

Why should you care? Distributed System Integration

Vision

Reality

Future

Vision

Reality

Future

UK e-Science Centres

Source: http://www.nesc.ac.uk/centres/

12

Why should you care?Vision

Reality

Future

Vision

Reality

Future

Source: “The Anatomy of the Grid”, Foster, Kesselman, Tuecke, 2001

The real and specific problem that underlies the Grid concept is coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations.

13

Why should you care?Terminology

Grid has strong links with “Utility Computing”, “Autonomic Computing” and “Service Oriented Architecture”.

Vision

Reality

Future

Vision

Reality

Future

14


Grid addresses pain points now [Reality]Grids are built not bought, but are delivering

real benefits in commercial settings Low utilization of enterprise resources High cost of provisioning for peak demand Inadequate resources prevent use of

advanced applications Lack of information integration

Vision

Reality

Future


15

Why should you care?Early Commercial Applications

Vision

Reality

Future


“Gridified” Infrastructure

FinancialServices

DerivativesAnalysis

Statistical Analysis

Portfolio Risk

Analysis

DerivativesAnalysis

Statistical Analysis

Portfolio Risk

Analysis

Manufacturing

Mechanical/ Electronic

Design

Process Simulation

FiniteElement Analysis

Failure Analysis

Mechanical/ Electronic

Design

Process Simulation

FiniteElement Analysis

Failure Analysis

LS / Bioinformatics

Cancer Research

Drug Discovery

Protein Folding

Protein Sequencing

Cancer Research

Drug Discovery

Protein Folding

Protein Sequencing

Other

Web Applications

Weather Analysis

Code Breaking/

Simulation

Academic

Web Applications

Weather Analysis

Code Breaking/

Simulation

Academic

Sources: IDC, 2000 and Bear Stearns- Internet 3.0 - 5/01 Analysis by SAI

Gri

d S

erv

ice

s M

ark

et

Op

po

rtu

nit

y 2

00

5

Energy

Seismic Analysis

Reservoir Analysis

Seismic Analysis

Reservoir Analysis

Entertainment

Digital Rendering

Digital Rendering

Massive Multi-Player

Games

Massive Multi-Player

Games

Streaming Media

Streaming Media

Leading adopters (Oct 2003) *• Financial services: 31%• Life sciences: 26%• Manufacturing: 18%

*Grids 2004: From Rocket Science To Business Service, The 451 Group

16

Why should you care?Grid Deployment Strategies

A range of excellent commercial & open source products for resource federation Federate enterprise computing resources Federate enterprise information resources Globus Toolkit®: inter-enterprise sharing

But, “Grids are built, not bought” Integration with other enterprise systems is

needed to deliver complete solution

Start small & with well-defined ROI case Grow based on experience

Vision

Reality

Future


17Image courtesy Christian Richters: Source:Wired News

Data Grids for High Energy Physics

Fastest particle accelarator: Large Hadron ColliderWhen completed in 2005, CERN's Large Hadron Collider will send protons and ions from hydrogen nuclei rushing through a 17-mile circular tunnel at speeds of up to 52,200,000 miles per hour.

18Image courtesy Harvey Newman, Caltech

Tier2 Centre ~1 TIPS

Online System

Offline Processor Farm

~20 TIPS

CERN Computer Centre

FermiLab ~4 TIPSFrance Regional Centre

Italy Regional Centre

Germany Regional Centre

InstituteInstituteInstituteInstitute ~0.25TIPS

Physicist workstations

~100 MBytes/sec

~100 MBytes/sec

~622 Mbits/sec

~1 MBytes/sec

There is a “bunch crossing” every 25 nsecs.

There are 100 “triggers” per second

Each triggered event is ~1 MByte in size

Physicists work on analysis “channels”.

Each institute will have ~10 physicists working on one or more channels; data for these channels should be cached by the institute server

Physics data cache

~PBytes/sec

~622 Mbits/sec or Air Freight (deprecated)




Caltech ~1 TIPS

~622 Mbits/sec

Tier 0Tier 0

Tier 1Tier 1

Tier 2Tier 2

Tier 4Tier 4

1 TIPS is approximately 25,000

SpecInt95 equivalents

Data Grids for High Energy Physics

19

Mathematicians Solve NUG30Quadratic Assignment Problem

MetaNEOS: Argonne, Iowa, Northwestern, WisconsinSource:Shawn McKee The Grid:The Future of High Energy Physics Computing? January 7,2002

Location 1

Location 2

Location 3

Location 4

The distances are:•d(1,2) = 22, •d(1,3) = 53, •d(2,3) = 40, •d(3,4) = 55.

The required flows between facilities are:

•f(2,4) = 1, •f(1,4) = 2, •f(1,2) = 3, •f(3,4) = 4.

The permutation p corresponding to this graphical solution is ( 2, 1, 4, 3 ).

20

Mathematicians Solve NUG30Looking for the solution to the NUG30 quadratic assignment problem An informal collaboration of mathematicians and computer scientistsCondor-G delivered 3.46E8 CPU seconds in 7 days (peak 1009 processors) in U.S. and Italy (8 sites)

NUG30 Solution: 14,5,28,24,1,3,16,15,10,9,21,2,4,29,25,22,13,26,17,30,6,20,19,8,18,7,27,12,11,23

MetaNEOS: Argonne, Iowa, Northwestern, WisconsinSource:Shawn McKee The Grid:The Future of High Energy Physics Computing? January 7,2002

21

Network for Earthquake Engineering Simulation NEESgrid: national infrastructure to couple earthquake engineers with experimental facilities, databases, computers, & each otherOn-demand access to experiments, data streams, computing, archives, collaboration

NEESgrid: Argonne, Michigan, NCSA, UIUC, USC

22

The 13.6 TF TeraGrid: Computing at 40 Gb/s

26

24

8

4 HPSS

5

HPSS

HPSS UniTree

External Networks

External Networks

External Networks

External Networks

Site Resources Site Resources

Site ResourcesSite ResourcesNCSA/PACI8 TF240 TB

SDSC4.1 TF225 TB

Caltech Argonne

TeraGrid/DTF: NCSA, SDSC, Caltech, Argonne www.teragrid.org

23U.S. PIs: Avery, Foster, Gardner, Newman, Szalay www.ivdgl.orgImage courtesy of http://www.sdss.org/news/releases/20050111.yardstick.html

iVDGL:International Virtual Data Grid Laboratory

Sloan Digital Sky Survey is the most ambitious astronomical survey project ever undertaken. The survey will map in detail one-quarter of the entire sky, determining the positions and absolute brightnesses of more than 100 million celestial objects. It will also measure the distances to more than a million galaxies and quasars

24U.S. PIs: Avery, Foster, Gardner, Newman, Szalay www.ivdgl.org

iVDGL:International Virtual Data Grid Laboratory

Tier0/1 facility

Tier2 facility

10 Gbps link

2.5 Gbps link

622 Mbps link

Other link

Tier3 facility

25


An open Grid is to your advantage [Future] Standards are being defined now that will determine the future of this technology

Vision

Reality

Future


26

Grid Vision, Marketing, and Reality

Vision Computing & data resources can be shared

like content on the Wb

Marketing Have we got a [Data, compute, knowledge,

information, desktop, PC, enterprise, cluster, …] Grid for you!

Reality Commercial products mostly noninteroperable Open source tools offer de facto standards,

but are also far from a complete solution

Vision

Reality

Future


27

Standards Matter!Open, standard protocols Enable interoperability Avoid product/vendor lock-in Enable innovation/competition on end points Enable ubiquity

In Grid space, must address how we Describe, discover, & access resources Monitor, manage, & coordinate, resources Account & charge for resources

For many different types of resource

Vision

Reality

Future


28

Open Grid Services Architecture

Define a service-oriented architecture … the key to effective virtualization

… that addresses vital “Grid” requirements AKA utility, on-demand, system management,

collaborative computing in particular, distributed service management

… building on Web services standards extending those standards where needed

“The Physiology of the Grid: An Open Grid Services Architecture for Distributed Systems Integration”, Foster, Kesselman, Nick, Tuecke, 2002

Vision

Reality

Future

29

A family of six Web services specifications A design pattern to

specify how to use Web services to access “stateful” components

Message-based publish-subscribe to Web services

Latest Step Forward:WS-Resource Framework

Gro

ups

References

Noti

fica

tion

Faults

Properties

Lifetim

e

Vision

Reality

Future


30

WS-Resource Framework Completes Grid-WS Convergence

Grid

Web

The definition of WSRF means that Grid and Web communities can move forward on a common base

WSRF

Started far apart in apps & tech

OGSI

GT2

GT1

HTTPWSDL,

WS-*

WSDL 2,

WSDM

Have beenconverging

Vision

Reality

Future


31

The Evolution of the GRID1980’s Parallel computing clusters -

improved performance from tightly coupled clusters and data sharing

1990’s Grid 1: Extend the advances in parallel computing to geographically distributed systems

2000 Grid II: Grid is a platform for integrating loosely coupled applications: some components running in parallel and some for linking disparate resources largely developed in the serial-von-Neumann paradigm - storage, visualisation, a-d/d-a converters and sensors

32

The Evolution of the GRID

Currently there are (clusters) of very powerful computing/ communications systems (i) Systems for acquiring digital data and processing data (Amazon.com or Oracle clusters)

(ii) Systems for analysing and visualising information (CERN’s large hadron collider, Protein Synthesis systems)

(iii) Systems for imaging, analysis and visualisation for distributed data (weather prediction, satellite based military civilian systems)

(iv) Systems that can link Sensors and predict on real-time information (military systems, video surveillance)

33

The Evolution of the GRIDDevelopments in networking technologies, operating systems, clustered data bases, application services and device technologies have enabled developers to build systems with literally distributed millions of nodes for providing: • Web-based services personal commercial transactions

• Content delivery networks that can cache web-pages seamlessly

• Wireless networks have spawned ad-hoc distributed systems that when linked to wide-area networks lead to a complex distributed system.

Problems of efficiency, reliability, accessibility and security are not addressed in ‘global’ terms.

34


* Sputnik

1960 1970 1975 1980 1985 1990 1995 2000

* ARPANET

* Email* Ethernet

* TCP/IP* IETF

* Internet Era * WWW Era

* Mosaic

* XML

* PC Clusters* Crays * MPPs

* Mainframes

* HTML

* W3C

* P2P

* Grids

* XEROX PARC wormCO

MP

UTIN

GC

om

mu

nic

ati

on

* Web Services

* Minicomputers

* PCs

* WS Clusters

* PDAs* Workstations

* HTC

Source: www.gridbus.org

35

2100

2100 2100 2100 2100

2100 2100 2100 2100

Personal Device SMPs or SuperComputers

LocalCluster

GlobalGrid

PERFORMANCE

+

Q

o

S

•Individual•Group•Department•Campus•State•National•Globe•Inter Planet•Universe

Administrative Barriers

EnterpriseCluster/Grid


Source: www.gridbus.org

36


Grid is being developed not only to make distributed resources available to end-user not also to co-ordinate such usage for sharing and aggregation of resources.

37


Moore’s law improvements in computing produce highly functional end-systemsThe internet and burgeoning wired and wireless provide wide-spread connectivityChanging modes of working and problem solving emphasise teamwork, computationNetwork growth produce dramatic changes in topology and geography

38


The first generation involved proprietary solutions for sharing high-performance computing resourcesThe second generation introduced middleware to cope with scale and heterogeneityThe third generation introduced a service-oriented approach leading to commercial projects in addition to the scientific projects now collectively known as e-Science

39

The Evolution of the GRIDThe first generation

FAFNER, I-WAY

The second generation Technologies: Globus, Legion Distributed object systems (Jini and RMI, The

common component architecture form) Grid resource brokers and schedulers Grid portals Integrated systems Peer-to-Peer computing

The third generation Service-oriented architecture (web services, OGSA,

Agents) Information aspects: relation with the World Wide

Web Live information systems

40

The Evolution of the GRIDGlobus Toolkit® History

0

5000

10000

15000

20000

25000

30000

1997 1998 1999 2000 2001 2002

Do

wn

loa

ds

pe

r M

on

th f

rom

ftp

.glo

bu

s.o

rg

DARPA, NSF, and DOE begin funding Grid work

NASA beginsfunding Grid work,DOE adds support

The Grid: Blueprint for a New Computing

Infrastructure published

GT 1.0.0Released

Early ApplicationSuccesses Reported

NSF & European CommissionInitiate Many New Grid Projects

Anatomy of the GridPaper Released Significant

CommercialInterest inGrids

Physiology of the GridPaper Released

GT 2.0Released

Does not include downloads from:NMI, UK eScience, EU Datagrid,IBM, Platform, etc.

Source: Ian Foster’ s presentation on “The First 50 Years” , British Computer Society, Lovelace Medal Award Presentation, May, 2003

41

Building blocks of the Grid

NetworksComputational ‘nodes’ on the GridPulling it all togetherCommon infrastructure: standards

42

GRID: Key Issues

Resources Discovery, Allocation, Scheduling

Availability

Access, Security, Networks

Efficiency Economy, Management Administration.

Hardware Computers, Services, Networks

Application

Development, Testing

43

GRID: Key Issues Sharing

A biochemist will be able to exploit 10,000 computers to screen 100,000 compounds in an hour1,000 physicists worldwide will be able to pool resources for petop analyses of petabytes of data

A multidisciplinary analysis in aerospace couples code and data in geographically distributed organisations may be possibleCivil engineers colloborate to design, execute, and analyse shake table experiments

Climate scientists will be able to visualise, annotate, and analyse terabyte simulation datasets

44DOE X-ray grand challenge: ANL, USC/ISI, NIST, U.Chicago

tomographic reconstruction

real-timecollection

wide-areadissemination

desktop & VR clients with shared controls

Advanced Photon Source

GRID: Key Issues Sharing Online Access to Scientific Instruments

archival storage

45

MORE DEFINTIONS

ResourceNetwork protocolNetwork enabled serviceApplication Programming Interface(API)Software Development Kit (SDK)Syntax

46

MORE DEFINTIONS: Resource

An entity that is to be shared E.g., computers, storage, data, software

Does not have to be physical entity E.g., Condor pool, distributed file system,…

Defined in terms of interfaces, not devices E.g. scheduler such as LSF and PBS define a

compute resource Open/close/read/write define access to a

distributed file system, e.g NFS, AFS, DFS

47

MORE DEFINTIONS: Network protocol

A formal description of message formats and a set of rules for message exchange Rules may define sequence of message

exchanges Protocol may define state-change in endpoint,

e.g. file system state change

Good protocols designed to do one thing Protocols can be layered

Examples of protocols IP, TCP, TLS( was SSL), HTTP, Kerberos

48

MORE DEFINTIONS: Network enabled services

Implementation of a protocol that defines a set of capabilities Protocol defines interaction with

service All services require protocols Not all protocols are used to provide

services (e.g. IP, TLS)

Examples: FTP and Web servers

49

MORE DEFINTIONS: Application Programming Interface

(API)

A specification for a set of routines to facilitate application developmentSpec often language specific (or IDL) Routine name, number, order and type of

arguments; mapping to language constructs Behaviour or function of routine

Examples GSS API(security), MPI (message passing)

50

MORE DEFINTIONS: Software Development Kit (SDK)

A particular instantiation of APISDK consists of libraries and tools Provides implementation of API

specification

Can have multiple SDKs for an APIExamples of SDKs MPICH, Motif Widgets

51

MORE DEFINTIONS: Syntax

Rules for encoding information, e.g. XML, Condor ClassAds, Globus RSL

Distinct from protocols One syntax may be used by many

protocols

Syntaxes may be layered E.g., Condor ClassAds -> XML->ASCII

52

References

Berman F., Fox G., Hey T. (2003) Grid Computing: Making the Global Infrastructure a Reality, Chichester, John Willey & Sons Inc. http://www.computing.surrey.ac.uk/courses/csm23/list.html

53

CSM23 Assessment and WeightingComponents of Assessment

Method(s) Percentage weighting

Annotated bibliography Students are required to write a 200 word summary of each of 5 key research papers

10%

Oral Examination 20%

Laboratory Exercise Students are required to implement small-scale laboratory homework during the semester.

20%

Project Students are expected to implement a Grid project ad write IEEE formatted report about their projects. In addition, the students are asked to give a presentation.

Implementation:20%IEEE Report:20%Presentation:10%

54

CSM23 TimetableDate Topic Lecturer

17/01/2005 Overview and Motivation Mrs.Tugba Taskaya-Temizel

24/01/2005 Grid Architecture, Technologies and Resource Allocation

Mrs.Tugba Taskaya-Temizel

31/01/2005 Peer to Peer Computing Dr Nick Antonopoulos

7/02/2005 Parallel Computing

Security

Dr.Roger M A Peel, Dr.James Heather

14/02/2005 Parallel Computing

Security

Dr.Roger M A Peel, Dr.James Heather

21/02/2005 Grid Applications Mrs.Tugba Taskaya-Temizel

28/02,7/03, 14/03, 21/03

Seminars Mrs.Tugba Taskaya-Temizel

Lecture slides / notes.

Technology

grid technology

utility grid

electricity grid

terminology grid

grid deployment strategies

grid addresses pain

resources source

enterprise computing