Download Internet Computing and the Emerging Grid

An Overview Grid Computing
and Applications
Subject Code: 433-498
Rajkumar Buyya
Grid Computing and Distributed Systems (GRIDS) Lab.
The University of Melbourne
Melbourne, Australia
www.gridbus.org
WW Grid
Overview
Computing platforms and how the Grid is
different ?
 Towards global (Grid) computing.
 Grid resource management and scheduling.
 Application development challenges.
 Approaches to Grid computing.
Grid applications
Grid Projects in GRIDS Lab@ Melbourne
 Summary and conclusions

COMPUTING
* HTC
* Mainframes * Minicomputers
NETWORKING
Technologies Introduced
Major Networking and Computing
Technologies Introduction
1960
* PCs
* Crays
* XEROX PARC worm
* Email
* MPPs
* IETF
* Internet Era
* ARPANET
1970
* TCP/IP
* Ethernet
1975
1980
* PDAs
* Workstations
1985
* P2P
* Grids
* PC Clusters
* WS Clusters
* W3C
* HTML * Mosaic
* WWW Era
1990
1995
* Web Services
* XML
2000
Internet: Past, Present, Future
Number of hosts
(millions)
140
120
100
The 'Network Effect’
kicks in, and the web
goes critical'
80
60
40
20
0
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
TCP/IP
HTML
Mosaic
XML
PHASE
1. Packet Switching Networks
1969: 4 US Universities linked to form ARPANET
1972: First e-mail program created
1976: Robert Metcalfe develops Ethernet
2. The Internet is Born
TCP/IP becomes core protocol
Domain Name System created
IETF created (1986)
3. The World Wide Web
HTML hypertext system created
CERN launch World Wide Web
NCSA launch Mosaic interface
4. with XML
5. The Grid
Internet and WWW Growth
10,000,000
1,000,000
Internet Hosts
100,000
10,000
1,000
WWW Servers
100
10
4
1
1969
1970
1975
1980
1985
1990
1995
2000
Installed base and Growth rate for telephone
lines, mobile phones, & Internet hosts - 1995
Income Group/
Region
Installed, 1995
Phone Mobile Internet
Lines
Phones Hosts
1994-95 Growth Rates (%)
Phone Mobile Internet
Lines
Phones Hosts
Lower Income
2.0
0.12
1.35
35.7
135.1
246.0
Lower- Middle
9.1
0.33
73.31
8.7
105.1
167.0
Upper - Middle
14.5
1.34
380.13
6.4
66.8
111.9
High
53.2
8.70
10749.23
3.6
55.6
97.0
Africa
1.7
0.09
69.14
7.9
60.5
81.4
Americas
29.0
5.17
8359.58
5.4
42.3
91.5
Asia
5.4
0.62
121.70
14.7
108.3
150.0
Europe
33.0
3.04
2732.24
3.6
59.5
112.2
Oceans
39.7
9.55
12845.55
4.0
85.7
88.8
World
12.1
1.56
1661.89
7.0
60.4
97.8
Source: ACM, Nov, 97 (phones, international telecommunication union, hosts, network Wizards
Internet as a delivery Vehicle
Scalable HPC: Breaking
Administrative Barriers
2100
2100
2100
2100
2100
2100
2100
2100
?
P
E
R
F
O
R
M
A
N
C
E
2100
Administrative Barriers
•Individual
•Group
•Department
•Campus
•State
•National
•Globe
•Inter Planet
•Universe
Desktop
SMPs or
SuperComputers
Local
Cluster
Enterprise
Cluster/Grid
Global
Cluster/Grid
Inter Planet
Cluster/Grid ??
Why Grids ? Large Scale Exploration
needs them—Killer Applications.

Solving grand challenge applications using
computer modeling, simulation and analysis
Aerospace
Internet &
Ecommerce
Life Sciences
CAD/CAM
Digital Biology
Military Applications
Cluster of Clusters - Hyperclusters
Cluster 1
Scheduler
Master
Daemon
LAN/WAN
Submit
Graphical
Control
Cluster 3
Execution
Daemon
Scheduler
Clients
Master
Daemon
Cluster 2
Submit
Graphical
Control
Scheduler
Master
Daemon
Clients
Submit
Graphical
Control
Clients
Execution
Daemon
Execution
Daemon
Grid: Towards Internet Computing for
(Coordinated) Resource Sharing
http://www.sun.com/hpc/
Grid enables:
Resource
Sharing
Selection
Aggreation
- Unification of geographically distributed resources
What is Grid ?

A paradigm/infrastructure that
enabling the sharing, selection, & aggregation
of geographically distributed resources:


Wide
area



Computers – PCs, workstations, clusters, supercomputers,
laptops, notebooks, mobile devices, PDA, etc;
Software – e.g., ASPs renting expensive special purpose applications
on demand;
Catalogued data and databases – e.g. transparent access to
human genome database;
Special devices/instruments – e.g., radio telescope – SETI@Home
searching for life in galaxy.
People/collaborators.
[depending on their availability, capability, cost, and
user QoS requirements]
for solving large-scale problems/applications.
P2P/Grid Applications-Drivers

Distributed HPC (Supercomputing):


High-Capacity/Throughput Computing:


Medical instrumentation & Mission Critical.
Collaborative Computing:


Drug Design, Particle Physics, Stock Prediction...
On-demand, realtime computing:


Application service provides (ASPs) & Web services.
Data-intensive computing:


Sharing digital contents among peers (e.g., Napster)
Remote software access/renting services:


Large scale simulation/chip design & parameter studies.
Content Sharing (free or paid)


Computational science.
Collaborative design, Data exploration, education.
Service Oriented Computing (SOC):

Computing as Competitive Utility: New paradigm, new industries,
and new business.
Building and Using Grids requires...





Services that make our systems Grid Ready!
Security mechanisms that permit resources to
be accessed only by authorized users.
(New) programming tools that make our
applications Grid Ready!.
Tools that can translate the requirements of
an application into requirements for
computers, networks, and storage.
Tools that perform resource discovery,
trading, composition, scheduling and
distribution of jobs and collects results.
A Typical Grid Computing
Environment
Grid Information Service
Grid Resource Broker
R2
R3
R5
Application
database
R4
RN
Grid Resource Broker
R6
Grid Information Service
R1
Resource Broker
Issues in Grid Technology
Development
Sources of Complexity in Resource
Management for World Wide Computing










Size (large number of nodes, providers, consumers)
Heterogeneity of resources (PCs, Workstatations, clusters, and
supercomputers)
Heterogeneity of fabric management systems (single system image
OS, queuing systems, etc.)
Heterogeneity of fabric management polices
Heterogeneity of applications (scientific, engineering, and
commerce)
Heterogeneity of application requirements (CPU, I/O, memory,
and/or network intensive)
Heterogeneity in demand patters
Geographic distribution and different time zones
Differing goals (producers and consumers have different objectives
and strategies)
Unsecure and Unreliable environment
Traditional approaches to resource
management are NOT useful for Grid ?

They use centralised policy that need



Due to too many heterogenous parameters in the Grid it is
impossible to define:



complete state-information and
common fabric management policy or decentralised consensus-based
policy.
system-wide performance matrix and
common fabric management policy that is acceptable to all.
So, we propose the usage of “economics” paradigm for managing
resources





proved successful in managing decentralization and heterogeneity that
is present in human economies!
We can easy leverage proven Economic principles and techniques
Easy to regulate demand and supply
User-centric, scalable, adaptable, value-driven costing, etc.
Offers incentive (money?) for being part of the grid!
Grid Resource Management
systems need to ensure/provide:


Site autonomy.
Heterogeneous resources and substrate:





Each resource can be different – SMPs, Clusters,
Linux, UNIX, Windows, Intel, etc.
Resource owners have their own policies or scheduling
mechanisms (Codine/Condor).
Extend policies, through resource brokers.
Resource allocation/co-allocation
Online control - can apps (Graphics) tolerate nonavailability of a resource and adapt themselves?
Grid RMS to support
•Authentication (once).
•Specify (code, resources,
etc.).
•Discover resources.
authorization,
•Negotiate authorisation,
acceptable
acceptableuse,
use,Cost,
Cost,etc.
etc.
•Acquire resources.
Jobs.
•Schedule jobs.
•Initiate computation.
Domain 1
Domain 2
•Steer computation.
•Access remote data-sets.
•Collaborate with results.
•Account for usage.
Ack: Globus..
Resource Management Architecture
Resource Brokers
(RSL Specialization)
RSL
Application
Resource Co-allocators
Local Resource Mgr Local Resource Mgr
Information
Service - MDS
Local Resource Mgr
Major Grid Projects
and Initiatives
mix-and-match
Object-oriented
Internet/partial-P2P
Network enabled Solvers
Economy/Service-Oriented Grid
Computing
Many Grid Projects & Initiatives

Australia







Nimrod-G
GridSim
Virtual Lab
Gridbus
DISCWorld
..new coming up








UNICORE
MOL
UK eScience
Poland MC Broker
EU Data Grid
EuroGrid
MetaMPI
Dutch DAS
XW, JaWS
Japan


Ninf
DataFarm
Korea...
N*Grid
USA






Europe









Cycle Stealing & .com Initiatives




Globus
Legion
OGSA
Javelin
AppLeS
NASA IPG
Condor-G
Jxta
NetSolve
AccessGrid
and many more...
Distributed.net
SETI@Home, ….
Entropia, UD, Parabon,….
Public Forums




Global Grid Forum
P2P Working Group
IEEE TFCC
Grid & CCGrid conferences
http://www.gridcomputing.com
Initiative
Focus and Technologies Developed
UNICORE
The UNiform Interface to Computer Resources aims to deliver software that allows users to submit
jobs to remote high performance computing resources – www.fz-juelich.de/unicore
MOL
Metacomputer OnLine is a toolbox for the coordinated use of WAN/LAN connected systems. MOL aims
at utilizing multiple WAN-connected high performance systems for solving large-scale problems that
are intractable on a single supercomputer – www.uni-paderborn.de/pc2/projects/mol
METODIS
Metacomputing Tools for Distributed Systems –
www.hlrs.de/structure/organisation/par/projects/metodis/
Globe
Globe is a research project aiming to study and implement a powerful unifying paradigm for the
construction of large-scale wide area distributed systems: distributed shared objects –
www.cs.vu.nl/~steen/globe
Pozan
Poznan Centre works on development of tools and methods for metacomputing www.man.poznan.pl/metacomputing/
Date Grid
This project aims to develop middleware and tools necessary for the data-intensive applications of
high-energy physics – grid.web.cern.ch/grid
MetaMPI
MetaMPI supports the coupling of heterogeneous MPI systems, thus allowing parallel applications
developed using MPI to be run on Grids without alteration – www.lfbs.rwthaachen.de/~martin/MetaMPICH/
DAS
This is a wide-area distributed cluster, used for research on parallel and distributed computing by five
Dutch universities – www.cs.vu.nl/das
JaWs
JaWS is an economy-based computing model where both resource owners and programs using these
resources place bids to a central marketplace that generates leases of use – roadrunner.ics.forth.gr
Initiative
Focus and Technologies Developed
Globus
This project is developing basic software infrastructure for computations that integrate geographically
distributed computational and information resources – www.globus.org
Legion
Legion is an object-based metasystem. Legion supports transparent scheduling, data management, fault
tolerance, site autonomy, and a wide range of security options – www.legion.virginia.edu
Javelin
Javelin: Internet-based parallel computing using Java – www.cs.ucsb.edu/research/javelin/
AppLes
This is an application-specific approach to scheduling individual parallel applications on production
heterogeneous systems – www.infospheres.caltech.edu/
NASA IPG
The Information Power Grid is a testbed that provides access to a Grid – a widely distributed network of high
performance computers, stored data, instruments, and collaboration environments – www.ipg.nasa.gov
Condor
This project aims is to develop, deploy, and evaluate mechanisms and policies that support high throughput
computing (HTC) on large collections of distributed computing resources – www.cs.wisc.edu/condor/
Harness
Harness builds on the concept of the virtual machine and explores dynamic capabilities beyond what PVM can
supply. It focused on developing three key capabilities: Parallel plug-ins, Peer-to-peer distributed control, and
multiple virtual machines – www.epm.ornl.org/harness
NetSolve
NetSolve is a project that aims to bring together disparate computational resources connected by computer
networks. It is a RPC based client/agent/server system that allows one to remotely access both hardware and
software components – www.cs.utk.edu/netsolve/
Grid Port
SDSCs Grid Port Toolkit generalises the HotPage infrastructure to develop a reusable portal toolkit –
gridport.npaci.edu/
HotPage
NPACI’s HotPage is a user portal that is designed to be a single point-of-access to computer resources –
hotpage.npaci.edu/
Gateway
Gateway offers a programming paradigm implemented over a virtual Web of accessible resources www.npac.syr.edu/users/haupt/WebFlow/demo.html
Initiative
Focus and Technologies Developed
Ninf
Ninf allows users to access computational resources including
hardware, software and scientific data distributed across a wide
area network with an easy-to-use interface – ninf.etl.go.jp
Bricks
Bricks is a performance evaluation system that allows analysis
and comparison of various scheduling schemes on a typical highperformance global computing setting – matsuwww.is.titech.ac.jp/~takefusa/bricks
Initiative
Focus and Technologies Developed
DISCWorld An infrastructure for service-based metacomputing across LAN
and WAN clusters. It allows remote users to login to this
environment over the Web and request access to data, and also
to invoke services or operations on the available data –
dhpc.adelaide.edu.au/Projects/DISCWorld/
Nimrod/G A global scheduler (resource broker) for parametric computing
& GRACE over clusters or computational grids –
www.dgs.monash.edu.au/~rajkumar/ecogrid
Many Testbeds ? & who pays ?
GUSTO
EcoGrid
Legion Testbed
NASA IPG
Some GRID
APPLICATIONS
Types of Grid Applications
Sequential – dusty deck codes.
 Data Parallel:




Asynchronous:



Synchronous – tightly coupled;
Loosely synchronous.
Irregular in time and space;
Difficult to parallelise to exploit the massive
parallelism.
Embarrassingly Parallel.
Grid Applications-Drivers

Distributed HPC (Supercomputing):


High-throughput computing:


Data mining, particle physics (CERN), Drug Design.
On-demand computing:


Application service provides (ASPs).
Data-intensive computing:


Sharing digital contents among peers (e.g., Napster)
Remote software access/renting services:


Large scale simulation/chip design & parameter studies.
Content Sharing


Computational science.
Medical instrumentation & network-enabled solvers.
Collaborative:

Collaborative design, data exploration, education.
Distributed Supercomputing
(SF-Express/MPICH-G, Caltech)
NCSA
Origin
Caltech
Exemplar


CEWES
SP
Maui
SP

SF-Express distributed
interactive simulation.
100K vehicles (2002 goal)
using 13 computers, 1386
nodes, 9 sites.
Globus mechanisms for




Resource allocation;
Distributed startup;
I/O and configuration;
Security.
P. Messina et al., Caltech
http://www.globus.org/applications/
SF-Express Architecture




MPI and socket
communication;
Hand startup.
Interest
Mgmt.
Local
Simulation
Router

Create synthetic,
representations of
interactive environments.
Scalability via interest
management.
Starting point:
Router
Local
Simulation
Router
Interest
Mgmt.
Interest
Mgmt.
Local
Simulation
High Throughput Computing
(parameter sweep applications)





A study involving exploration of possible scenarios i.e., execution of the same program for various design
alternatives (data).
It consists of large number of tasks (1000s).
Generally, no inter-task communication (task farming).
Large size data (MBytes+) files and I/O constraints
A large class of application areas:



Parameter explorations and simulations (Monte Carlo);
A large number of science, engineering, and commercial
applications: Astrophysics, Drug Design, NeroScience, Network
simulation, structural engineering, automobiles crash simulation,
aerospace modeling, financial risk analysis
Condor, Nimrod/G, DesignDrug@Home, SETI@Home,
FOLD@Home, Distributed.net.
Ad Hoc Mobile Network Simulation
Ad Hoc Mobile Network Simulation: Network performance under
different microware frequencies and different weather conditions –
uses Nimrod.
Drug Design: Data Intensive
Computing on Grid
Molecules
Protein

Chemical Databases
(legacy, in .MOL2 format)
It involves screening
millions of chemical
compounds (molecules) in
the Chemical DataBase
(CDB) to identify those
having potential to serve as
drug candidates.
DesignDrug@Home Architecture
A Virtual Lab for “Molecular Modeling for Drug Design” on P2P Grid
Data Replica
Catalogue
Grid Market
Directory
“Give me list PDBs sources
Of type aldrich_300?”
“Screen 2K molecules
in 30min. for $10”
Grid Info.
Service
GTS
Resource
Broker
“mol.5 please?”
GTS
(RB maps suitable
Grid nodes and
Protein DataBank)
PDB2
GTS
GTS
PDB1
GTS
(GTS - Grid
Trade Server)
MEG(MagnetoEncephaloGraphy)
Data Analysis on the Grid: Brain Activity
Analysis
64 sensors MEG
2
Analysis All pairs (64x64) of MEG data by
shifting the temporal region of MEG data
over time: 0 to 29750: 64x64x29750 jobs
Data Generation
3
1
5
Results
Data Analysis
Nimrod-G
4
Life-electronics laboratory,
AIST
•Provision of expertise in
the analysis of brain function
•Provision of MEG analysis
•[deadline, budget, optimization preference]
World-Wide Grid
[Collaboration with Osaka University, Japan]
SETI@home: Search for Extraterrestrial
Intelligence at Home
Content Sharing – P2P
Collaborative Engineering
Access GRID: http://www-fp.mcs.anl.gov/fl/accessgrid/
Components of an AG Node
RGB Video
Digital Video
NETWORK
Digital Video
Digital Audio
Control
Computer
Rick Stevens & Team, ANL
Display
Computer
Video
Capture
Computer
Audio
Capture
Computer
NTSC Video
Analog Audio
Mixer
Echo
Canceller
•Group to group interactions.
•Human collaboration across
distributed locations
•Remote visualizations, virtual meeting,
seminars,etc.
•Uses grid technologies for secure
communication etc.
•May have interaction with scientific apps.
Image-Rendering
http://www.swin.edu.au/astronomy/pbourke/povray/parallel
Parallelisation of Image Rendering


Image splitting (by rows, columns, and
checker)
Each segment can be concurrently
processed on different nodes and render
image as segments are processed.
Scheduling (need load balancing)


Each row rendering
takes different times
depending on image
nature – e.g,
rendering rows across
the sky take less time
compared to those
that intersect the
interesting parts of
the image.
Rending apps can be
implemented using
MPI, PVM, or p-study
tools like Nimrod and
schedule.
Data Intensive Computing
e.g., CERN Data Grid initiative
CERN Large Hadron Collider - circular particle accelerator
to be placed in 27 km long tunnel in 2005.
Conclude with a comparison
with the Electrical Grid………..
Where we are ????
Alessandro Volta in Paris in 1801 inside
French National Institute shows the battery
while in the presence of Napoleon I
Fresco by N. Cianfanelli (1841)
(Zoological Section "La Specula" of National History Museum of Florence University)
What ?!?!
Oh, mon
Dieu !
This is a mad man…
….and in the future,
I imagine a
worldwide
Power (Electrical)
Grid …...
2000 - 1801 = 199 Years
What will be the dominant Grid approach in the next future ??
”The Computational Grid” is
analogous to Electricity (Power)
Grid and the vision is to offer a
(almost) dependable, consistent,
pervasive, and inexpensive
access to high-end resources
irrespective their location of
physical existence and the
location of access.
Trends
It is very difficult to predict the future and
this is particular true in a field such as
Information Technology
“I think there is a world market for about five computers.”
Thomas J. Watson Sr., IBM Founder, 1943
Trends
Grid
The time is exciting but the way
ahead may be hard and long….!
The Grid Impact!
“The global computational grid is
expected to drive the economy of the
21st century similar to the electric
power grid that drove the economy of
the 20th century”
Future Grid Scenarios






Access to any resources, for anyone, anywhere,
anytime, from any platform – portal (super)
computing .
Application access to resources from the wall
socket!
Many applications provide solutions in realtime.
Choice of working: office vs home vs . . .
Collaboratories for distributed teams.
Monitoring and steering applications through
wireless devices (PDAs etc.).
Final Summary


There are currently a large number of
projects and diverse range of emerging
Grid developmental approaches being
pursued.
These range from metacomputing
frameworks to application testbeds, and
from collaborative environments to batch
submission mechanisms.
Conclusions





The HPC will be dominated by Peer-to-Peer
Grid of clusters.
Adaptive, scalable, and easy to use Systems
and End-User applications will be prominent.
Access electricity, internet, entertainment
(music, movie,…), etc. from the wall socket!
An Economics –based Service Oriented Grid
Computing computing needed for eventual
success of Grids!
The impact of Grid on 21st century economy will
be the same as electricity on 20th century
economy.
Further Information

Books:



IEEE Task Force on Cluster Computing


High Performance Cluster Computing, V1, V2,
R.Buyya (Ed), Prentice Hall, 1999.
The GRID, I. Foster and C. Kesselman (Eds),
Morgan-Kaufmann, 1999.
http://www.ieeetfcc.org
GRID Forums

www.gridforum.org, www.egrid.org
CCGRID 2001, www.ccgrid.org
 GRID Meeting - www.gridcomputing.org

Further Information

Cluster Computing Infoware:


Grid Computing Infoware:


http://www.gridcomputing.com
IEEE DS Online - Grid Computing area:


http://www.buyya.com/cluster/
http://computer.org/dsonline/gc
Millennium Compute Power Grid/Market
Project

http://www.ComputePower.com
Thank You… Any ??