clustering by akashmshah
DESCRIPTION
THANKS FOR VISIT ON MY PRESENTATION.TRANSCRIPT
![Page 1: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/1.jpg)
CLUSTER COMPUTING
PREPARED BY:- AKASH M SHAH &KEYUR AADESARA (5TH SEM CE)
![Page 2: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/2.jpg)
4
OVERVIEWHigh performance computing
High throughput computing
HPC, and HTC Parallel algorithms
Software technologies
![Page 3: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/3.jpg)
CLUSTER COMPONENTS
1. Multiple High Performance Computers
2. Operating Systems.
A.Linux B. Microsoft NT
3. High Performance Networks.
4. Network Interface Card.
5. Parallel Programming Environments &tools.
parallelprocessingwasperformedbymultipleprocessorsinaspeciallydesignedparallelcomputer.Thesearesystemsinwhichmultipleprocessorsshareasinglememoryandbusinterfacewithinasinglecomputer.Thesetypesofclusterincreaseavailability,performance,andscalabilityforapplications,particularlycomputationallyordataintensivetasks.
Thebasicbuildingblocksofclustersarebrokendownintomultiplecategories:1.ClusterNodes2.ClusterNetwork3.NetworkCharacterizationTherearethreeprimarycategoriesofapplicationsthatuseparallelclusters:1.ComputeIntensiveApplication.2.DataorI/OIntensiveApplications.3.TransactionIntensiveApplications.Ausersubmitsajobtotheheadnode.Thejobidentifiestheapplicationtorunonthecluster.Thejobschedulerontheheadnodeassignseachtaskdefinedbythejobtoanodeandthenstartseachapplicationinstanceontheassignednode.Resultsfromeachoftheapplicationinstancesarereturnedtotheclientviafilesordatabases.
Themainbenefitsofclustersare:1.Availability2.Performance3.ScalabilityThesebenefitsmaptoneedsoftoday'senterprisebusiness,education,militaryandscientificcommunityinfrastructures.
![Page 4: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/4.jpg)
CONTENTS
Introduction Architecture History Application Advantages Disadvantages
![Page 5: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/5.jpg)
INTRODUCTIONWhat is cluster computing?
cluster computing is the technique of linking two or more computers into a network (usually through a local area network) inorder to take advantage of the Parallel processing power of those computers.
![Page 6: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/6.jpg)
INTRODUCTION
The clustering model can provide both HA & HP, and also manageability, scalability, & affordability
Typically homogeneous, tightly coupled, nodes trust each other.
As number of h/w components rises, so does the probability of failure.
Increasing probability of fault occurrence for long-running applications.
![Page 7: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/7.jpg)
WHY WE USE CLUSTER COMPUTING????
Improve the operating speed ofprocessors & other components.
Connect multiple processors together & coordinate their computational efforts.
allow the sharing of a computational task among multiple processors
![Page 8: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/8.jpg)
How does it works?A user submits a job to the head node.
The job identifies the application to run on the cluster.
The job scheduler on the head node assigns each task defined by the job to a node and then starts each application instance on the assigned node.
Results from each of the application instances are returned to the client via files or databases.
![Page 9: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/9.jpg)
CLUSTER COMPUTING ARCHITECTURE
![Page 10: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/10.jpg)
12
HISTORYCluster computing system contract by IBM in the 1950s based on the MIT Whirlwind computer architecture.During the decade of the 1980s, increased interest in the potential of cluster computing was marked by important experiments in research and industry.
![Page 11: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/11.jpg)
TYPES OF CLUSTERSHigh Availability Clusters
Load-balancing Clusters
High-performance Clusters
![Page 12: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/12.jpg)
HIGH AVAILABILITY CLUSTER
The clusters are designed to maintain redundant nodes that can act as backup systems in the event of failure. The minimum number of nodes in a HA cluster is two - one active and one redundant - though most HA clusters will use considerably more nodes.
![Page 13: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/13.jpg)
High Availability Clusters
HIGH AVAILABILITY CLUSTER
![Page 14: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/14.jpg)
LOAD-BALANCING CLUSTER
Load-balancing clusters are extremely useful for those working with limited IT budgets. Load-balancing clusters operate by routing all work through one or more load-balancing front-end nodes.
![Page 15: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/15.jpg)
HIGH-PERFORMANCE CLUSTER
HPC clusters are designed to exploit parallel processing power of multiple nodes.
![Page 16: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/16.jpg)
18
NETWORKEthernet
10 Mbps obsolete100 Mbps almost obsolete1000 Mbps standard
Protocols
TCP/IP
![Page 17: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/17.jpg)
19
Example Clusters
July 19991000 nodes Used for genetic algorithm research by John Koza, Stanford Universitywww.genetic-programming.com/
Mateti-Everything-About-Linux
![Page 18: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/18.jpg)
![Page 19: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/19.jpg)
Mateti-Everything-About-Linux 21
Largest Cluster SystemIBM BlueGene, 2007Memory: 73728 GBInterconnect: ProprietaryPowerPC 440106,496 nodes478.2 Tera FLOPS
![Page 20: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/20.jpg)
APPLICATION
There are three primary categories of applications that use parallel clusters.
1.Compute intensive application.
2.Data or i/o intensive applicaion.
3.Transaction intensive application.
![Page 21: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/21.jpg)
ADVANTAGESHigh Performance
Expandability
Scalability: the resources are distributed in case of Cluster Computing.
Easy to upgrade and maintain.
High Availability
![Page 22: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/22.jpg)
DISADVANTAGES
Software: difficult to develop software for distributed systems.
Network:- saturation, transmissions.
Security: easy access also applies to secrete data. Increased power: consumption compare to single server.
![Page 23: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/23.jpg)
CONCLUSIONSolve parallel processing paradox .
Clusters based supercomputers can be seen everywhere!
New trends in hardware and software tech-nologies are likely to make clusters more promising and fill SSI (Single System Image)gap.
![Page 24: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/24.jpg)
FUTURE TRENDS - GRID COMPUTING
The Grid is a large system of computing resources that performs tasks and provides to users a single point of access, commonly based on the World Wide Web interface, to these distributed resources.
Major Grid projects include NASA’s Information Power Grid, two NSF Grid projects (NCSA Alliance’s Virtual Machine Room ).
http://hello-engineers.blogspot.com/
![Page 25: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/25.jpg)
More Information on Clusters
http://www.ieeetfcc.org/
http://lcic.org/
http://beowulf.org/
http://linuxclusters.com/
Sun Microsystems - http://www.sun.com/
![Page 26: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/26.jpg)
![Page 27: Clustering by AKASHMSHAH](https://reader036.vdocuments.site/reader036/viewer/2022062616/5495d8beb47959525e8b462a/html5/thumbnails/27.jpg)
REFRERENCES