Post on 11-May-2015
description
11-12-2013
1
GRID COMPUTING
Reference : Internet
SPEAKER’S BACKGROUND IN GRID AND CLOUD
COMPUTING
�An opening caveat
�This talk is based on my:
� My doctoral research in Grid Computing and
Shared memory
� Review and session chair experience for
various conferences
� Expert talks at various Institutes for faculty
and students in STTPs, Conferences etc.
� Experience as a Project mentor
12/1
1/2
013
2
11-12-2013
2
KEY TAKEAWAYS
�Widen personal technical spectrum
� Understand the evolution of recent
technologies
� Envision the upcoming technological trends
� Be useful in Teaching STTP related and allied
subjects
� Guiding projects –UG/PG
12/1
1/2
013
3
ELECTRICAL GRID
12/1
1/2
013
4
11-12-2013
3
WHY GRIDS ? LARGE SCALE EXPLORATION NEEDS
THEM—KILLER APPLICATIONS.
� Solving grand challenge applications using computer modeling,
simulation and analysis
Life Sciences
CAD/CAM
Aerospace
Military ApplicationsDigital Biology
Military ApplicationsMilitary Applications
Internet &
Ecommerce
12/1
1/2
013
5
6
WHAT IS A GRID?
�Early defs: Foster and Kesselman, 1998
“A computational grid is a hardware and software
infrastructure that provides dependable,
consistent, pervasive, and inexpensive access to
high-end computational facilities”
�Kleinrock 1969:
“We will probably see the spread of ‘computer
utilities’, which, like present electric and
telephone utilities, will service individual homes
and offices across the country.”
12/1
1/2
013
11-12-2013
4
GRID COMPUTING
12/1
1/2
013
7
ELEMENTS OF GRID COMPUTING
� Resource sharing
� Computers, data, storage, sensors, networks, …
� Sharing always conditional: issues of trust, policy,
negotiation, payment, …
� Coordinated problem solving
� Beyond client-server: distributed data analysis,
computation, collaboration, …
� Dynamic, multi-institutional virtual
organizations
� Community overlays on classic org structures
� Large or small, static or dynamic
12/1
1/2
013
8
11-12-2013
5
9
12/
11/
201
3
A Typical Grid Computing Environment
Grid Resource Broker
Resource Broker
Application
Grid Information Service
Grid Resource Broker
databaseR2R3
RN
R1
R4
R5
R6
Grid Information Service
2
GRID COMPUTING: ATTRIBUTES (1)
� Virtualization: Abstracting grid entity into service
� It enables grid components to integrate tightly without creating rigidity and brittleness in the system.
� Components quickly react to change, and adapt failures without compromising performance and reliability.
� Dynamic Provisioning:
� It simply means distributing supplies where they are needed. Supplies mean request, data, computation
� A grid service broker links grid elements together automatically and dynamically, based on the knowledge about their requirements and attributes. and adjust the association according to the change and failures
12/1
1/2
013
10
11-12-2013
6
GRID COMPUTING: ATTRIBUTES (2)
� Resource Pooling: contributes to lower cost
� Consolidation and pooling is used for better utilization.
� Provide flexibility to optimize the association
� Self-Adaptive Software:
� Everyday task of administrator are automated and simplified. The bulk of maintenance and tuning is automated to reduce IT staff cost.
� Unified Management:
� Self-adaptive software does not eliminate human interaction. Unified management is provided to simplify the management process.
� Single tool can be used to provision, monitor, administer
12/1
1/2
013
11
HOW ARE GRIDS USED?
High-performance computing
Collaborative data-sharing
Collaborative design
Drug discovery
Financial modeling
Data center automation
High-energy physics
Life sciences
E-Business
E-Science
BELLE
Natural language
processing
Utility computing
Business Intelligence
(Data Mining)
12/1
1/2
013
12
11-12-2013
7
12/11/2013 13DOE X-ray grand challenge: USC NIST, U.Chicago
tomographic reconstruction
real-time
collection
wide-area
dissemination
desktop & VR clients with shared controls
Advanced Photon Source
ONLINE ACCESS TO
SCIENTIFIC INSTRUMENTS
archival
storage
GRID COMPUTING: COMPONENTS
• Computation:– Computing cycles provided by processors of grid machines.
Simple, parallel, iterative uses of computing elements
• Networked Storage:– Integrated view of data storage (Datagrid). Local disk,
secondary storage, mountable, Unified name space
• Network interconnects:– Fast interconnection technologies. Redundant and external
Internet connections makes parallel processing faster, and management better.
• Software and licenses– Expensive software, Sharing Expensive licenses, Limited
use of multiple installation
12/1
1/2
013
14
11-12-2013
8
GRID ARCHITECTURE (SERVICE-ORIENTED)
Applications
Grid Architected Services
Web Services (Extended Web Services)
Servers NetworkStorage
Security MessagingDatabaseFilesystems Directory
12/1
1/2
013
15
GRID USERS GRID TYPES
16
� Grid developers
� Tool developers
� Application developers
� End Users
� System
Administrators
� Compute grid
� Data grid
� Grids on the Internet
12/1
1/2
013
11-12-2013
9
OPEN GRID SERVICES ARCHITECTURE
� Developed by the Global Grid Forum to define a
common, standard, and open architectures for
Grid-based applications.
� Provides a standard approach to all services on the
Grid.
� VO Management Service.
� Resource discovery and management service:
� Job management service.
� Security services.
� Data management services.
� Built on top of and extends the Web Services
architecture, protocols, and interfaces.
12/1
1/2
013
17
18
CDAC: GARUDA COMPONENT ARCHITECTURE
12/1
1/2
013
11-12-2013
10
MANY GRID PROJECTS & INITIATIVES
� Australia
� Nimrod-G
� GridSim
� Virtual Lab
� Gridbus
� DISCWorld
� ..new coming up
� Europe
� UNICORE
� MOL
� UK eScience
� Poland MC Broker
� EU Data Grid
� EuroGrid
� MetaMPI
� Dutch DAS
� XW, JaWSJapan
� Ninf
� DataFarm
� Korea...
N*Grid
� USA
� Globus
� Legion
� OGSA
� Javelin
� AppLeS
� NASA IPG
� Condor-G
� Jxta
� NetSolve
� AccessGrid
� and many more...
� Cycle Stealing & .com Initiatives
� Distributed.net
� SETI@Home, ….
� Entropia, UD, Parabon,….
� Public Forums
� Global Grid Forum
� P2P Working Group
� IEEE TFCC
� Grid & CCGrid conferences
http://www.gridcomputing.com
12/1
1/2
013
19
GRID APPLICATIONS-DRIVERS
� Distributed HPC (Supercomputing):
� Computational science.
� High-throughput computing:
� Large scale simulation/chip design & parameter studies.
� Content Sharing
� Sharing digital contents among peers (e.g., Napster)
� Remote software access/renting services:
� Application service provides (ASPs).
� Data-intensive computing:
� Data mining, particle physics (CERN), Drug Design.
� On-demand computing:
� Medical instrumentation & network-enabled solvers.
� Collaborative:
� Collaborative design, data exploration, education.
12/1
1/2
013
20
11-12-2013
11
GLOBUS
Five parts:
�Common Runtime� GT Core for building new services
�Security� To provide secure access. Based upon Grid Security
Infrastructure (GSI)
�Execution management� Initiation, monitoring, management, scheduling and
coordination of executable programs (jobs)
�Data management� Discover, transfer, and access large data
� Information services� Discover & monitor dynamic services
12/1
1/2
013
21
22
GLOBUS TOOLKIT
12/1
1/2
013
11-12-2013
12
Data Management
SecurityCommonRuntime
Execution Management
Information Services
Web Services
Components
Non-WS
Components
Pre-WSAuthenticationAuthorization
GridFTP
GridResource
Allocation Mgmt(Pre-WS GRAM)
Monitoring& DiscoverySystem(MDS2)
C CommonLibraries
GT2
WSAuthenticationAuthorization
ReliableFile
Transfer
OGSA-DAI[Tech Preview]
GridResource
Allocation Mgmt(WS GRAM)
Monitoring& DiscoverySystem(MDS4)
Java WS Core
CommunityAuthorization
ServiceGT3
ReplicaLocationService
XIO
GT3
CredentialManagement
GT4
Python WS Core[contribution]
C WS Core
CommunitySchedulerFramework
[contribution]
DelegationService
GT4
Globus Open Source Grid Software 12/1
1/2
013
23
Data Management
SecurityCommonRuntime
Execution Management
Information Services
Web Services
Components
Non-WS
Components
Pre-WSAuthenticationAuthorization
GridFTP
GridResource
Allocation Mgmt(Pre-WS GRAM)
Monitoring& DiscoverySystem(MDS2)
C CommonLibraries
GT2
WSAuthenticationAuthorization
ReliableFile
Transfer
OGSA-DAI[Tech Preview]
GridResource
Allocation Mgmt(WS GRAM)
Monitoring& DiscoverySystem(MDS4)
Java WS Core
CommunityAuthorization
ServiceGT3
ReplicaLocationService
XIO
GT3
CredentialManagement
GT4
Python WS Core[contribution]
C WS Core
CommunitySchedulerFramework
[contribution]
DelegationService
GT4
Globus Open Source Grid Software 12/1
1/2
013
24
11-12-2013
13
GRID SIMULATOR: OPTORSIM
� Models the interactions of the individual components of a
running Data Grid.
� Simulates optimization and replication
12/1
1/2
013
25
CONVERGENCE OF COMPETING
PARADIGMS/COMMUNITIES NEEDED
� Web
� Data Centres
� Utility Computing
� Service Computing
� Grid Computing
� P2P Computing
� Cloud Computing
� Market-Oriented Computing
� …
•Ubiquitous
access
•Reliability
•Scalability
•Autonomic
•Dynamic
discovery
•Composability
•QoS
•SLA
•…
} +
Paradigms
Attributes/Capabilities
?-Trillion $ business
- Who will own it?
Manjrasoft
12/1
1/2
013
26
11-12-2013
14
VOLUNTEERS ARE SELDOM PAID; NOT
BECAUSE THEY ARE WORTHLESS, BUT
BECAUSE THEY ARE PRICELESS!
--AUTHOR UNKNOWN
Contact:
principal@vit.edu.in; 4.seema@gmail.com
+919833818846