The Grid: past, present, future

Sandeep Kumar PooniaHead Of Dept. CS/IT

B.E., M.Tech., UGC-NET

LM-IAENG, LM-IACSIT,LM-CSTA, LM-AIRCC, LM-SCIEI, AM-UACEE

The Grid

• The Grid is the computing and datamanagement infrastructure that will providethe electronic underpinning for a globalsociety in business, government, research,science and entertainment

“Grid infrastructure will provide us with the ability todynamically link together resources as anensemble to support the execution of large-scale,resource-intensive, and distributed applications.”

• Grids, integrate networking, communication, computation andinformation to provide a virtual platform for computation and datamanagement in the same way that the Internet integratesresources to form a virtual platform for information.

A COMMUNITY GRID MODEL

Global Resources

• The bottom horizontal layer of the CommunityGrid Model consists of the hardware resourcesthat underlie the Grid. Such resources includecomputers, networks, data archives,instruments, visualization devices and so on.

Common Infrastructure

• The next horizontal layer consists of the softwareservices and systems which virtualizes the Grid.

• Community efforts such as NSF’s Middleware Initiative(NMI), OGSA, as well as emerging de facto standardssuch as Globus provide a commonly agreed upon layerin which the Grid’s heterogeneous and dynamicresource pool can be accessed.

• The key concept at the common infrastructure layer iscommunity agreement on software, which willrepresent the Grid as a unified virtual platform andprovide the target for more focused software andapplications.

User and Application-Focused Grid Middleware, Tools and Services

• The next horizontal layer contains softwarepackages built atop the common infrastructure.

• This software serves to enable applications tomore productively use Grid resources by maskingsome of the complexity involved in systemactivities such as authentication, file transfer, andso on.

• Portals, community codes, application schedulingsoftware and so on reside in this layer andprovide middleware that connects applicationsand users with the common Grid infrastructure.

Grid applications

• The topmost horizontal layer representsapplications and users.

• The Grid will ultimately be only as successfulas its user community and all of the otherhorizontal layers must ensure that the Gridpresents a robust, stable, usable and usefulcomputational and data managementplatform to the user.

New Devices – Sensors, PDAs, and Wireless.

• The vertical layers represent the next steps forthe development of the Grid.

• The vertical layer on the left represents theinfluence of new devices – sensors, PDAs, andwireless.

• Over the next 10 years, these and other newdevices will need to be integrated with the Gridand will exacerbate the challenges of managingheterogeneity and promoting performance.

Policies for Sharing and using Resources

• At the same time, the increasing globalization of theGrid will require serious consideration of policies forsharing and using resources, global-area networkingand the development of Grid economies.

• As we link together national Grids to form a GlobalGrid, it will be increasingly important to develop Gridsocial and economic policies which ensure the stabilityof the system, promote the performance of the usersand successfully integrate disparate political,technological and application cultures.

BUILDING BLOCKS OF THE GRID

Networks• The heart of any Grid is its network – networks link

together geographically distributed resources andallow them to be used collectively to support executionof a single application.

• If the networks provide ‘big pipes’, successfulapplications can use distributed resources in a moreintegrated and data-intensive fashion;

• if the networks provide ‘small pipes’, successfulapplications are likely to exhibit minimalcommunication and data transfer between programcomponents and/or be able to tolerate high latency.

The Internet2 Abilene network in the US

In 2002, such national networks exhibit roughly 10 Gbps backbone performance.

UK National Backbone Research and Education Network

APAN Asian Network

International Networks

Although there are exceptions, one can capture a typical leadingGrid research environment as a 10 : 1 : 0.1 Gbps ratio representing:

national: organization: desktop links

Today, new national networks are beginning to change this ratio.The GTRN or Global Terabit Research Network initiative linknational networks in Asia, the Americas and Europe with aperformance similar to that of their backbones.

By 2006, GTRN aims at a 1000 : 1000 : 100 : 10 : 1 gigabitperformance ratio representing:

International backbone: National: Organization: Optical Desktop: Desktop links

This implies a performance increase of over a factor of 2 per year innetwork performance, and clearly surpasses expected CPUperformance and memory size increases of Moore’s law

Common Infrastructure: Standards

The development of key standards that allow thecomplexity of the Grid to be managed by softwaredevelopers and users without heroic efforts is critical tothe success of the Grid.

Both the Internet and the IETF , and the Web and theW3C consortium have defined key standards such asTCP/IP, HTTP, SOAP, XML and now WSDL – the Webservices definition language that underlines OGSA.

Such standards have been critical for progress in thesecommunities.

The GGF is now building key Grid-specific standardssuch as OGSA, the emerging de facto standard for Gridinfrastructure.

Common Infrastructure: Standards

In addition, NMI and the UK’s Grid Core Program areseeking to extend, standardize and make more robustkey pieces of software for the Grid arsenal such asGlobus, Condor, OGSA-DAI and the Network WeatherService.

In the last two decades, the development of PVM andMPI, which pre-dated the modern Grid vision,introduced parallel and distributed computingconcepts to an entire community and provided theseeds for the community collaboration, whichcharacterizes the Grid community today.

GRID APPLICATIONS AND APPLICATIONMIDDLEWARE

Life science applications

One of the fastest-growing application areas inGrid Computing is the Life Sciences.

Computational biology, bioinformatics, genomics,computational neuroscience and other areas areembracing Grid technology as a way to access,collect and mine data [e.g. the Protein Data Bank,the myGrid Project , the Biomedical InformationResearch Network (BIRN)], accomplish large-scalesimulation and analysis (e.g. MCell), and toconnect to remote instruments

Biomedical Informatics Research Network – oneof the most exciting new applicationmodels for the Grid.

Engineering-oriented applications

The Grid has provided an important platform for making resourceintensive engineering applications more cost-effective. One of themost comprehensive approaches to deploying production Gridinfrastructure and developing large-scale engineering-oriented Gridapplications is the NASA IPG in the United States. The NASA IPGvision provides a blueprint for revolutionizing the way in whichNASA executes large-scale science and engineering problems viathe development of

persistent Grid infrastructure supporting ‘highly capable’ computingand data management services that, on demand, will locate and co-schedule the multicenter resources needed to address large-scaleand/or widely distributed problems,

ancillary services needed to support the workflow managementframeworks that coordinate the processes of distributed scienceand engineering problems.

A Grid for aerospace engineering showinglinkage of geographically separated subsystemsneeded by an aircraft.

Data-oriented applications

Data is emerging as the ‘killer application’ ofthe Grid.

Over the next decade, data will come fromeverywhere – scientific instruments,experiments, sensors and sensornets, as wellas a plethora of new devices.

The Grid will be used to collect, store andanalyze data and information, as well as tosynthesize knowledge from data.

Distributed Aircraft Maintenance Environment (DAME) Grid tomanage data from aircraft engine sensors.

Physical science applications

Physical science applications are another fast-growingclass of Grid applications.

Much has been written about the highly innovative andpioneering particle physics–dominated projects – the GriPhyN,

Particle Physics Data Grid, and

iVDGL projects in the United States and

the EU DataGrid,

the UK GridPP and

the INFN (Italian National Institute for Research in Nuclearand Subnuclear Physics) Grid projects

Architecture of particle physics analysis Grid

Commercial ApplicationsIn the commercial world, Grid, Web and distributed computing, and

information concepts are being used in an innovative way in a wide varietyof areas including inventory control, enterprise computing, games and soon.

Enterprise computing areas where the Grid approach can be applied include

end-to-end automation,end-to-end security,virtual server hosting,disaster recovery,heterogeneous workload management,end-to-end systems management,scalable clustering,accessing the infrastructure,‘utility’ computing,

accessing new capability more quickly,better performance,reducing up-front investment,gaining expertise not available internally, andWeb-based access (portal) for control (programming) of enterprise function.

Next-generation Grid applications

Next-generation Grid applications will includethe following:Adaptive applications (run where you can find

resources satisfying criteria X),

Real-time and on-demand applications (dosomething right now),

Coordinated applications (dynamic programming,branch and bound) and

Poly-applications (choice of resources for differentcomponents).