Introduction to Grid Computing

Introduction to Grid Computing

• 1.
  INTRODUCTION TO GRIDCOMPUTING-BY ABHIJEET NAWALB.E.(Honors), CSEMBM ENGG. COLLEGE(www.mbm.ac.in)
• 2.
  WHY GRID -PROBLEMDOMAINS:There are some large scale problems which require analysis of high volumes of data (of the order of petabytes) to derive an inference.
• 3.
  Such problems thusneed computing resources (say processing power of the order of teraflops, secure dependable networks and storage) often for which even a supercomputer does not suffice.
• 4.
  Some Such Problems:1) High Energy Physics Experiments (E.g. LHC) 2) Earthquake Simulation Experiments 3) Drug Discovery 4) Weather Modeling Grid aims to provide a cost effective, scalable and reliable computing platform to such grand challenge data intensive problems.WHAT IS A GRID:
• 5.
  GRIDS AND THEIRTYPES:Grid is an infrastructure that involves the integrated and collaborative use of computers, networks, databases and scientific instruments.
• 6.
  Grid resources maybe owned and managed by multiple organizations often heterogeneous (platforms, operating systems) and geographically dispersed in nature.
• 7.
  Computational Grid: Maincomputing resource shared – Processing Power.
• 8.
  Data Grid: Maincomputing resource shared – Storage. Provides fast and secure access to distributed, heterogeneous pools of data.
• 9.
  Network Grid: Mainpurpose to provide fault-tolerant and high performance communication. COMPUTING APPROACHES:
• 10.
  WHAT ARE GRIDREQUIREMENTS:Remote Storage and Replication of data sets.
• 11.
  Publication and discoveryof data sets using global logical name.
• 12.
  Security- access authorizationand authentication.
• 13.
  Discovery and Unifiedaccess and view of Remote Resources taking into account their local access policies.
• 14.
  Mapping and Schedulingof Jobs.
• 15.
  Movement of Code/Data between the user desktop machines and distributed resources.
• 16.
  Enforcement of Qualityof service requirements.
• 17.
  Metering and Accountingof resource usage.
• 18.
  Failure Detection andFailover Mechanisms.Virtual Organizations and Resource Sharing:Software Application ProviderWeatherPredictionProblem ADynamic Virtual Organization Aformed to provide weather predictionWeather predictionMath ModelingApplicationHardware Service ProviderUserComputerClusterVirtual organizations areLogical entities
• 19.
  Dynamically created tosolve a specific problem
• 20.
  On-demand resource allocationand provisioning for solving the problemNetworkBandwithBlades…resourcesUserFinancial Expert Service ProviderDynamic Virtual Organization Bformed to solve a financial modelingFinancialModelingProblem BFinancial ModelingDatabaseSystem
• 21.
  Virtual Organizations Continue..The concept of the VO is the key to grid computing.
• 22.
  A virtual organization(VO) is a dynamic group of individuals, groups, or organizations who define the conditions and rules (business objectives and policies) for sharing resources.
• 23.
  Although the resourcesin a grid might be in separate administrative domains, they are made available as virtual local resources to any node on the grid.
• 24.
  Dynamically collect resourcesfrom different providers based upon the users’ needs or urgency and the priority of the problem.
• 25.
  Identify problems, linkingthem to required resources or service providers, and automatically resolve them.
• 26.
  Dynamically provision andmanage the resources meeting QoS (Quality Of Service) while not violating SLAs and local autonomy of resource usage.GRID ARCHITECTURE:
• 27.
  GRID COMPUTING OPERTIONALVIEW:END USER APPLICATION INTERFACE1GRID MARKETDIRECTORY42Results and Cost Info.6. JobsGRID RESOURCE BROKERRESOURCES7. Results35GRID INFORMATION SERVICEACCOUNTING AND METERING SYSTEM
• 28.
  OPERATIONAL VIEW CONTINUE..1.User interacts with the Grid System through the Interface.
• 29.
  2. Submission ofthe user’s Quality of Service (QoS) and Computing needs to Grid Resource Broker (GRB)
• 30.
  3. GRB discoversresources and characteristics using Grid Information Service (GIS).
• 31.
  4. GRB identifiesresource prices from Grid Market Directory (GMD)
• 32.
  5. GRB ensuresuser’s credit and authorization privileges.
• 33.
  6. GRB’S schedulermaps and deploys data analysis jobs to QoS meeting resources.
• 34.
  7. GRB collatesthe results and account usage and cost information to the user. WHAT ARE THE BENEFITS:Technically:
• 35.
  Exploiting under utilizedresources
• 36.
  Access to additionalresources
• 37.
  Resource Balancing
• 38.
  Parallel CPU capacity
• 39.
  Scalability
• 40.
  Socially:
• 41.
  Resources can beharnessed for the good of society. For eg grids for drug discovery.
• 42.
  Commercially:
• 43.
  An opportunity tocreate a large processing center and sell its capabilities to customers.
• 44.
  This frees thecustomers from the need to purchase, configure, and manage their own large network of computers. Where large computing jobs are infrequent enough that renting is less expensive than purchasing computers. Eg: multi-player gaming grid –SECOND LIFE.GRID COMPUTING: FUTUREGRID INFRASTRUCTURE DEVELOPMENT
• 45.
  GRID MIDDLEWARE RESEARCH
• 46.
  STANDARIZATION -to allowinteroperability between different products and implementations.Open Grid Service Infrastructure (OGSI) standard .WSRF (Web Services Resource Framework) standard that enables Web services integration within Grid Architecture. This allows Grid Developers to take advantage of standard message formats and communication mechanisms such as HTTP and XML for communicating between heterogeneous components and architectures.
• 47.
  REFERENCES:“Weaving computational grids: How analogous are they with Electrical grids” by MadhuChetty and RajkumarBuyya.
• 48.
  “A gentle introductionto grid computing and technologies’- RajkumarBuyya and SrikumarVenugopal.
• 49.
  “Grid computing- abrief technology analysis”- Roger Smith.
• 50.
  “Introduction to GridComputing”- IBM Redbooks.THANK YOU