Grid Computing
Mansoor AhmedAziz 7865Warda Aqil
BSE/BCE – 5
Abstract
There is no consensus among scholars and technologywriters about the origins of grid computing. Perhaps thedifferent opinions are due to terminology and that someauthors use utility computing while others prefer gridcomputing.
Key Words
-
Grid, Discipline, Field, Subject, Subtopicand.
1. Introduction
Grid computing (or the use of a computational grid) isapplying the resources of many computers in a network to a single problem at the same time - usually to ascientific or technical problem that requires a greatnumber of computer processing cycles or
access to largeamounts of data. It is a form of distributed computing based on the dynamic sharing of resources between participants, organizations and companies with the aimof combining these resources and carrying out intensivecomputing applications or the processing of vast amountsof data. The creation of a "virtual supercomputer" byusing a network of geographically dispersed computers.It also refers to the automated sharing and coordinationof the collective processing power of many widelyscattered, robust computers that are not normallycentrally controlled, and that are subject to openstandards. Other terms employed in this context include:Autonomic computing, Data-centre virtualization, On-demand computing, Public resource computing, andutility computing. It also can be used for load balancingas well as high availability by employing multiplecomputers.
2. History
There is no consensus among scholars and technologywriters about the origins of grid computing. Perhaps thedifferent opinions are due to terminology and that someauthors use utility computing while others prefer gridcomputing. It's probably fair to say that the origins of grid computing came out of the early days of computer networks where using the "spare" CPU cycles was seenas an efficient and cost-effective way of getting the mostof what was then very expensive hardware.Its 1970: Computers are big mainframes that costhundreds of thousands of dollars, every second has to beaccounted for, and those otherwise "wasted" cycles can beused to get the most out of the cost.The term Grid computing originated in the early 1990sas a metaphor for making computer power as easy toaccess as an electric power grid in Ian Foster and CarlKesselmans seminal work, "The Grid: Blueprint for anew computing infrastructure".CPU scavenging and volunteer computing were popularized beginning in 1997 by distributed.net andlater in 1999 by SETI@home to harness the power of networked PCs worldwide, in order to solve CPU-intensive research problems.
3. The Grid Problem
Grid computing has evolved into an importantdiscipline within the computer industry by differentiatingitself from distributed computing through an increasedfocus on resource sharing, co-ordination, manageability,and high performance. The focus on resource sharing iscalled
the grid problem
, which can be defined as the setof problems associated with resource sharing among a setof individuals or groups. This sharing of resources,ranging from simple file transfers to complex andcollaborative problem solving, is accomplished under controlled and well-defined conditions and policies. Inthis context, the critical problems are resource discovery,authentication, authorization, and access mechanisms.Resource sharing is further complicated when a grid isintroduced as a solution for utility computing,
wherecommercial applications and resources become availableas shareable and on-demand resources. This concept of commercial on-demand utility grid services adds new,more difficult challenges to the already complicated grid problem list, including service level features, accounting,usage metering, flexible pricing, federated security,scalability, and open-ended integration.
4. Evolution of Grid Computing
During recent years, we have witnessed a major paradigm shift in distributed computing principles, with
a focus towards service orientation, open standardsintegration, collaboration, and virtualization. One particular area of interest centers on the evolution of gridcomputing principles into the mainstream of distributedcomputing and Web services. In this paper, we focus our analysis on this evolution and the significance of achieving some form of standardization of grid-computing architecture principles. This paper presentsthe technology standards that are driving major gridinitiatives and explains in simple terms how thesestandards and technologies are aligned with the IBM ondemand business concepts. In addition, we discuss therecent Web services specifications related to state fullresources (i.e., resources whose behavior is defined withrespect to their underlying state) and how these standardsrelate to grid computing. We conclude with discussionsexploring major aspects of grid-computing adoptionmodels and some significant attributes that influence thetransformation, collaboration, and virtualization featuresof these models.
5. Grid as a Virtual Organization
A
virtual organization
(VO) is a dynamic group of individuals, groups, or organizations who define theconditions and rules for sharing resources. The conceptof the VO is the key to grid computing. All VOs sharesome characteristics and issues, including commonconcerns and requirements that may vary in size, scope,duration, sociology, and structure. The members of anyVO negotiate the sharing of resources based upon therules and conditions defined by the VO, and the membersthen share the resources in the VO's constructed resource pool.
6. Grid Computing Different fromDistributed Computing
Grid computing can be differentiated from almost alldistributed computing paradigms by this definingcharacteristic:
The essence of grid computing lies in theefficient and optimal utilization of a wide range of heterogeneous, loosely coupled resources in anorganization tied to sophisticated workload management capabilities or information virtualization.
(Anorganization can span multiple departments, physicallocations, and so on.)
7. Grid Architecture
A new architecture model and technology has beendeveloped for the establishment and management of cross-organizational resource sharing. This newarchitecture, called
grid architecture
, identifies the basiccomponents of a grid system. The grid architecturedefines the purpose and functions of its components,while indicating how these components interact with oneanother.
The main focus of the architecture is oninteroperability among resource providers and users inorder to establish the sharing relationships. Thisinteroperability, in turn, necessitates common protocolsat each layer of the architectural model, which leads tothe definition of a grid protocol architecture.
8. High Level Overview of Grid Computing
The most common description of
grid computing
includes an analogy to a power grid. When you plug anappliance or other object requiring electrical power into areceptacle, you expect that there is power of the correctvoltage available, but the actual source of that power isnot known. Your local utility company provides theinterface into a complex network of generators and power sources and provides you with (in most cases) anacceptable quality of service for your energy demands.Rather than each house or neighborhood having to obtainand maintain its own generator of electricity, the power grid infrastructure provides a virtual generator. Thegenerator is highly reliable and adapts to the power needsof the consumers based on their demand.The vision of grid computing is similar. Once the proper kind of infrastructure is in place, a user will have accessto a virtual computer that is reliable and adaptable to theuser’s needs. This virtual computer will consist of manydiverse computing resources. But these individualresources will not be visible to the user, just as theconsumer of electric power is unaware of how their electricity is being generated. To reach this vision, theremust be standards for grid computing that will allow asecure and robust infrastructure to be built. Standardssuch as the Open Grid Services Architecture (OGSA) andtools such as those provided by the Globus Toolkit provide the necessary framework. Initially, businesseswill build their own infrastructures (what we might callintra-grids), but over time, these grids will becomeinterconnected. This interconnection will be made possible by standards such as OGSA and the analogy of grid computing to the power grid will become real.
9. Types of Grids
Grid computing can be used in a variety of ways toaddress various kinds of application requirements. Often,grids are categorized by the type of solutions that they best address. The three primary types of grids aresummarized below. Of course, there are no hard boundaries between these grid types and often grids may be a combination of two or more of these. However, asyou consider developing applications that may run in agrid environment, remember that the type of gridenvironment that you will be using will affect many of your decisions.
9.1. Computational Grid
A
computational grid
is focused on setting asideresources specifically for computing power. In this typeof grid, most of the machines are high-performanceservers.
9.2. Scavenging Grid
A
scavenging grid
is most commonly used with largenumbers of desktop machines. Machines are scavengedfor available CPU cycles and other resources. Owners of the desktop machines are usually given control over when their resources are available to participate in thegrid.
9.3. Data Grid
A
data grid
is responsible for housing and providingaccess to data across multiple organizations. Users arenot concerned with where this data is located as long asthey have access to the data. For example, you may havetwo universities doing life science research, each withunique data. A data grid would allow them to share their data, manage the data, and manage security issues suchas who has access to what data.
10. Grid Infrastructure Components
These are the main components of a gridinfrastructure:
•
Security
. Security is an important considerationin Grid computing. Each grid resource mayhave different security policies that need to becomplied with. A single sign-on authenticationmethod is a necessity. A commonly agreed-uponmethod of negotiating authorization is alsoneeded.
•
Resource management
. When a job issubmitted, the grid resource manager isconcerned with assigning a resource to the job,monitoring its status, and returning its results.
•
Information services
. For the grid resourcemanager to make informed decisions onresource assignments, the grid resource manager needs to know what grid resources are available,as well as their capacities and currentutilization. This knowledge about the gridresources is maintained and provided by GridInformation Service (GIS), also known as theMonitoring and Discovery Service (MDS).
•
Data management
. Data management isconcerned with how jobs transfer data or accessshared storage.
•
Scheduling
-- Work must be scheduled acrossthe service providers to ensure that they are kept busy and are working at appropriate times and periods. A standardized method of describingthe grid service will help give structure to thisarea, as it will enable grid implementations tospecify how work needs to be scheduled.
•
Work unit management
-- Effective gridservices require management of the distributionof work units to ensure that the work is evenlyspread over the service providers. Without astandard way of advertising and managing this process, it's possible for some service providersto sit idle while others have massive work queues that take them inordinate amounts of time to process.
•
Dispatch management
-- The role of brokeringwork units and dispatching them to clients can be handled in myriad ways. Not having astandard method of doing so, however, restrictsthe service providers that can connect to andaccept units of work and also restricts the abilityof grid services users -- the requesters -- tosubmit the work.The advantages of the standardized approachare many, but they all boil down to the same basic advantage: the extension and expansion of the resources available for grid computing. Morespecifically, following the OGSI standard should provide the following benefits:
•
Interoperability
-- It should be possible to mixand match service provider components, anddispatch tracking systems and systemsmanagement. Most important, though, is thatgrid systems can be developed and designed in avariety of languages and within a variety of different platforms easily and efficiently. All of these factors should make it easier to dispatchwork to service providers and for service providers to find grid services and systems thatthey can attach to.
•
Increased capacity
-- With more platforms andenvironments supported and the ability to moreeasily publish the services available, it shouldlead to an increase in the available capacity.
•
Flexibility
-- With a wider range of clients andfor the clients a wider range of grids, thisincreased flexibility means that grid users canincrease their computing capacity. With thecomponents talking the same language, we canswitch between grid systems and jobs from theclient and server perspective. This will allowgrid users to reserve more capacity and allowclients a wider choice of projects to support.
of 00
Grid Computing
1,741 Reads
Info and Rating
| Category: | |
| Rating: | (1 Rating) |
| Upload Date: | 02/03/2009 |
| Copyright: | Attribution Non-commercial |
| Tags: |
Leave a Comment
can u send it