Introduction to Grid Computing

Introduction to Grid

Computing

What is the

GRID?

Grid Computing Technology

• Grid computing enables the

virtualization of distributed

computing and data resources

such as processing, network

bandwidth and storage capacity

to create a single system

image, granting users and

applications seamless access

to vast IT capabilities.

• Just as an Internet user views a

unified instance of content via

the Web,

a grid user essentially

sees a single, large virtual

The Application-Infrastructure Gap

Global Community

What Grid can really do?

•

Help you build a secure IT infrastructure

•

Remote job submission so you can

distribute your workload

•

Middleware and Programming API for

developing distributed services and

application that use them.

Grid Computing’s Potential

X Unified view of data and

computers Computers and data appear to be local

X Efficient access to large data sets Caching Replication

Attributes X Single sign-on, security X Policy-based resource sharing

Type of Grid

• Computational Grid

– Present a virtual pool of computing resources

• Data Grid

– Present a virtual pool of storages

• Access Grid

Technology

and Trends

Brief History of Grid Computing

• Generation 0:

No Grid only Internet and Web

• Generation 1:

Internet Computing

– Seti Home, Peer-to-peer

– Many experimental Grid and Meta computing System

• Generation 2:

Execution oriented Grid

– Use grid as batch supercomputing system – Basic set of services, no standardize access – Workable security based on X.509

• Generation 3:

Services Oriented Grid

– Distributed Services accessible globally – Based on Web Services Standard

What is

really

the Grid?

• Set of Computers, Servers, Clusters connected together with good network

• OS Linux / Windows on each machine (mostly linux) • Install Grid Middleware, mostly Globus

• Setup so middleware properly including certificate infrastructure

• Use tool on top

– Scheduler to distribute computing workload

• Gridified Application

– Use globus library or other lobrary such as MPI – Grid services

Components of the Grid

High Speed Network

Grid Middleware Grid Enable Application

Grid Middleware

• Key technology in Grid Computing

‹

Provide uniform, high-level access to a wide

range of resources (including networks)

‹Address interdomain issues of security, policy, etc.

‹Permit application-level management and monitoring of end-to-end performance

• Middleware-level and higher-level APIs library

and tools for programmer to implement

The Globus Project

• Project by

– Argonne National Laboratory

– University of Southern California's Information Sciences Institute

– University of Chicago – University of Edinburgh

– Swedish Center for Parallel Computers

• Leader

– Ian Foster and Carl Kesselman

• Produces open-source software that is a de facto grid standard

• Globus is now being take care of by Globus alliance

Grid Architecture

– Protocol and interface that provide access to computing resources such as CPU, storage

• Connectivity Layer

– Protocol for Grid-specific network transaction such as security GSI

• Resources Layer

– Protocol to access a single resources from application

• GRAM (Grid Resource Allocation Management)

• GridFTP ( data access) • Grid Resource Information

Service • Collective layer

– Protocol that manage and access

group of resources Fabric

Connectivity Resources

Collective Layer Application Layer

Grid Security Infrastructure

• Core part of grid system

• Extension of

– SSL(Secure Socket Layer)/TLS (Transport

Layer Security)

– X.509 Certificate

• Provides

– Single sign-on capability

How GSI works

User Proxy

User Proxy

GSI--Enable Enable GRAM Server GRAM Server Process Process MAP MAP Single

Single SignonSignon using Grid using Grid Certificate and Certificate and ID ID

Remote process creation request

Remote process creation request

Local id

Local id

GSI

GSI--Enable Enable GRAM Server GRAM Server Process Process MAP MAP Local id Local id Gate Keeper Gate Keeper User obtain User obtain

signed certificate from CA

signed certificate from CA

GRAM

RSL

RSL

RSL

RSL

RSL

RSL

Bridging the Gap:

Service-Oriented

Infrastructure

• Service-oriented applications

– Wrap applications as

services

– Compose applications

into workflows

• Service-oriented

infrastructure

– Provision physical

resources to support

application workloads

Service ServiceAppln Users

Workflows

Composition Invocation

Open Grid Service Architecture

(OGSA)

• Reference architecture for the next generation grid technology

• Merge Grid and Web Services technologies together • open standards for grid computing

– Sponsored by the GGF (organization modeled after IETF) – Primary working groups: OGSA and OGSI

– Many vendors involved: IBM, Sun, Oracle, AVAKI, UD, etc…

• (But, ANL and IBM seem to have the upper hand)

Globus as

Service-Oriented Infrastructure

monitoring

Computers Specialized Storage resource User Application User Application User Application IBM IB M GRAM GridFTP Host Env User Svc DAIS Database Tool Tool Reliable File Transfer MyProxy Host Env User Svc MDS-Index

Applying the

Grid

How to Develop Grid

Application

• Use grid as parallel super computer

– Using Grid-enable-MPI (MPICH-G2) to do parallel programming

• Mostly apply to technical computing

• Use Grid as a powerful server system

– Using Portal + Grid Scheduler + to Distributed Jobs – Load balance in scheduler

• Use grid to develop SOA application

Traditional Grid Enable

Application

Grid and cluster Middleware (MPI, Globus)

Task Task

Task Task Task

Service Oriented Grid

Potential Application

• E-commerce

– Infrastructure for B2B communication

• E-government

– Secure information exchange for government

operation.

• Smart card, e-citizen

• E-Military

Potential Application

• E-health care

– Secure information exchange for health care industry

• Thailand Population: 63 Million People • Public Hospital: 200 Hospitals

• Private Hospital: 337 Hospitals

• 30 Baths Project: 1,480 Baths per capitar • 27,000 Doctors country wide

– Massive storage and processing are needed for medical data

• Patient record • Diagnostic record

Potential Application

• E-entertainment

– Video Streaming, secure exchange of digital

media

– Distributed gaming server, secure

subscription of gaming services

• E-learning

– Access Grid

What industries are using grid

computing now?

• Automotive and aerospace

– for collaborative design and data-intensive testing

• Financial services

– for running long, complex scenarios and arriving at more accurate decisions

• Life sciences

– for analyzing and decoding strings of biological and chemical information

• Government

– for enabling seamless collaboration and agility in both civil and military departments and agencies

• Higher education

– for enabling advanced, data and compute intensive research.

An eBusiness Use of Globus:

SAP Demonstration @

GlobusWorld

• 3 Globus-enabled applns:

– CRM: Internet Pricing Configurator (IPC)

– CRM: Workforce Management (WFM)

– SCM: Advanced Planner & Optimizer (APO)

• Applications modified to:

– Adjust to varying demand & resources – Use Globus to discover

& provision resources

IPC Dispatcher IPC Server Request: Price Query Delegation of Request Response: Pricelist Depending on: - Time - Discount - Number of Items - …

Web Browsers / Batch Processes

(typically several thousand requests)

IPC Server 1 2 2 3

SAP AG R/3 Internet Pricing & Configurator (IPC)

ThaiGrid Drug Discovery Project

• Partners: KU/IBM

• Problem: From over 1000 active compounds available from Thai medicinal plants database, find a smallest set of compounds that has a

potential to be used as a drug

– Very compute intensive.

Several month of computing time

• Solution: Use grid to increase computing power to 10-100 times

Concept

Grid and cluster Middleware (MPI, Globus)

Task Task

Task Task Task

Scheduler MMJFS Gamess Gamess Service Gamess File Server Portal Portlet OG SA DA I Broker Server Registration Server Backend DB Molecular DB Grid Ftp

Grid based data assimilation

using GA and Remote Sensing

• Partners: KU/AIT • Problem:

– Very long computation time from months to years – Large amount of RS data

needed to be moved around and process

• Solution:

– Using Grid to harvest more computing power – Hide data assimilation

process behind the service

8/22/2006 TAM2005 37

Massive Rendering of Graphics

Images

• Partners: KU/Imagimax • Problem:

– Rendering massive amount of images for animation require a massive amount of computing power

• Solution:

– Grid based Rendering Support Middleware

based on NPACI ROCKS – Use grid to link multiple

clusters together – Portal to submit

Nimrod-G, Australia

Astrophysics _{Air Pollution}

Antenna Design

Airfoil Design

Circuit Design _{Monte Carlo} Computational Chemistry

Public Health Policy CFD

Cardiac Modelling

Climate

A Typical eScience Use of Globus:

Network for Earthquake Eng.

Simulation

Links instruments, data,

computers, people

Grids for Physics:

LHC Computing Grid

Grid2003 Æ Open Science Grid

¾ 30 sites (2100-2800 CPUs) & growing

¾ 400-1300 concurrent jobs

¾ 8 substantial applications + CS experiments

¾ Running since October 2003

Korea

系统架构 ：DartGrid的分层模型

“We’ve used GT3 to build the largest database grid system for Traditional Chinese Medicine, integrating

about 50 TCM-relevant databases.” (Zhejiang U.) VO的地址栏 语义浏览面板 本体论树 Dar t Gr i d 虚拟组 织资源浏览面板 Q3语义查询 显示面板 语义注册面板

Global Grid Forum

(www.ggf.org)

• The Global Grid Forum (GGF) is a

community-initiated forum researchers and practitioners working on "grid" technologies. • GGF's primary objective is to

– promote and support the development, deployment, and implementation of Grid technologies and applications

– creation and documentation of "best practices" -technical specifications, user experiences, and implementation guidelines

• GGF has many working group working on various grid standard

Conclusion

• Grid Computing is a high impact technology

technology that will change the world

• Grid can help making the use of IT infrastructure

more efficient by

– Provides uniform access , security – Virtualized the resources

• Grid technology is new still quickly changing

More information

• Globus Alliance

– www.globus.org

• Grid Computing Planet

–

http://www.gridcomputingplanet.com/

• Vendor sites

Thank

You