China's massive education grid.

In one of the largest implementations to date of Grid computing, IBM and China's Ministry of Education will use Grid technology to let universities across the country collaborate on research, and scientific and education projects.

The China Education and Research Grid is being launched with six universities. When complete, it will link more than 200,000 students and faculty members at nearly 100 hundred universities across China. When the first phase of the project is complete in 2005, the Grid will perform more than six teraflops, or trillions of calculations per second, and eventually be capable of more than 15 trillion calculations per second. That would make the Grid more powerful than all but the world's largest supercomputer, the 35-teraflop Earth Simulator in Yokoharma, Japan. The Grid relies on Web services technology in WebSphere--IBM's Internet infrastructure software--based on the Open Grid Services Architecture (OGSA) standards that began as a joint project of IBM and the open-source Glabus Project. A total of 49 IBM eServer xSeries running Linux have been deployed, and the Grid also includes six units of pSeries servers running AIX and IBM TotalStorage FAStT200 servers for storing data.

"WebSphere software helps a Grid gather untapped computing capabilities and functionalities and make it available to users across the Grid as needed," says George Wang, director of IBM China Software Development laboratory and IBM China Research and Development Laboratory.

The China Grid will simplify how students and researchers access education and computing resources across China. Universities will be connected to a common virtual hub that will automatically find the appropriate application resources, from life sciences research to video courses and e-learning.

China's university system will save on development costs, since each school can focus on its area of expertise--e-learning or life sciences, for example--and tap into other applications as needed via the Grid.

A bioinformatics project will be among the first projects to run on the Grid The Bioinformatics Grid system, a cooperative project owned by Key laboratory of Bioinformatics, China Ministry of Education and Tsinghua University, was used to help identify the SARS gene and analyze similarities between different strains of the SARS virus.

Peking University's Real Course application will be another early project, providing students with speedier access to video courses by distributing information through distributed servers. Also, the University of Hong Kong e-learning application will let students practice Mandarin through an integrated learning portal with an easy-to-use Web interface that can verify the pronunciation of Mandarin characters through voice recognition. Hong Kong residents grow up speaking Cantonese, but many began studying Mandarin after Hong Kong rejoined mainland China.

With the WebSphere-powered Grid, China's universities will be able to organize the vast computational and informational resources of the country's entire higher educational system into a centralized, Internet- based hub to perform a wide range of complex tasks instantaneously. A specific request--such as a complex protein-folding computation for infectious disease research--can be pushed onto the Grid, automatically seek out an application at another campus that knows how to handle the computation, and feed it back to the original computer. IBM Grid computing technology has been deployed at Peking University, South China University of Technology, Tsinghua University, the University of Hong Kong, Xi'an giaotong University and Sun Yat-sen University. Other universities involved in the first phase of the project are Huazhong University of Science and Technology, Northeast University, Shandong University, Shanghai Jiactong University and Southeast University.

IBM and the universities will establish a Grid Application United R & D Center for the research and development of open Grid architecture solutions based on open standards such as OGSA and Web services. IBM will work with China's Ministry of Education on implementation, application development and training, and also teach students how to develop applications for and manage the Grid. The Ministry of Education and universities involved in the Grid project also will have advanced access to new Grid technologies from IBM

What is grid computing.

Grid computing enables the virtualization of distributed computing and data resources such as processing, network bandwidth and storage capacity to create one system image, granting users and applications seamless access to vast IT capabilities. Just as an Internet user views a unified instance of content via the Web, a grid user essentially sees a single, large virtual computer.

At its core, grid computing is based on an open set of standards and protocols--e.g., Open Grid Services Architecture (OGSA)--that enable communication across heterogeneous, geographically dispersed environments. With grid computing, organizatons can optimize computing and data resources, pool them for large capacity workloads, share them across networks and enable collaboration.

Evolution, not revolution

Grid can be seen as the latest and most complete evolution of more familiar developments--such as distributed computing, the Web, peer-to-peer computing and virtualization technologies.

* Like the Web, grid computing keeps complexity hidden: multiple users enjoy a single, unified experience.

* Unlike the Web, which mainly enables communication, grid computing enables full collaboration toward common business goals.

* Like peer-to-poor, grid computing allows users to share files.

* Unlike peer-to-peer, grid computing allows many-to-many sharing-not only files but other resources as well.

* Like clusters and distributed computing, grids bring computing resources together.

* Unlike clusters and distributed computing, which need physical proximity and operating homogeneity, grids can be geographically distributed and heterogeneous.

* Like virtualization technologies, grid computing enables the virtualization of IT resources.

Unlike virtualization technologies, which virtualize a single system, grid computing enables the virtualization of vast and disparate IT resources.

Also because grid computing dovetails naturally with powerful developments in IBM like autonomic computing and e-business on demand, grid offers businesses a host of practical benefits for coping with--and taking advantage of--an on demand world.

Frequently asked questions

A range of grid-related questions frequently asked by IBMers and customers alike.

1.0 What is a grid?

2.0 What effect does grid have on users whose machines are being utilized for processing?

3.0 Is grid computing available today-or is it more of a future statement?

4.0 What industries are using grid computing now?

5.0 What are the possible benefits of a grid deployment?

6.0 What is IBM's relationship with grid computing?

7.0 Grid and e-business on demand: what's the connection?

8.0 Does IBM use grid computing in its own infrastructure?

9.0 If 1 want to learn more about IBM Grid Computing, what's the first step?

10.0 What does it take to build a grid?

111.0 What about security in grid environments?

What Is a grid?

All or some of a group of computers, servers and storage across an enterprise, virtualized as one large computing system. Because grids unleash latent power that, at any one time, is not being used, they ran give companies a huge gain in power, speed and collaboration, radically accelerating compute-intensive processes. Cost, meanwhile, can remain low, as grids can be built using existing infrastructure, helping to ensure optimal utilization of computing capabilities.

What effect does grid have on users whose machines are being utilized for processing?

Grids are designed to be seamless and transparent. A user whose desktop PC, say, is contributing processing power to the grid will experience no negative effects: the grid runs in the background, utilizing available resources when needed by the system. If the PC user decides to run an application that requires more processing power, the work currently being processed on that machine may be dynamically reallocated to another machine in the grid with available processing power.

Is grid computing available today--or is it more of a future statement?

Grid computing is used today by many companies across a number of industries. Current IBM customer references for grid include Butterfly.net, a development studio, online publisher and infrastructure provider for massively multiplayer games that connect players on PC's, consoles and mobile devices. Butterfly Grid consists of two clusters of approximately 50 eServer xSeries servers running in IBM hosting facilities. Specialized game servers and database servers are fully meshed over high-speed fiber-optic lines, enabling transparent routing of players to different servers in the grid. Another current reference for ISM Grid Computing is the University of Pennsylvania's groundbreaking National Digital Mammography Archive, which gives rapid retrieval of digital patient files from multiple locations in a secure environment. The University of Pennsylvania Grid manages this huge data volume, schedules traffic and encrypts all image and information transmission using portal systems running almost exclusively on ISM hardware--including sixteen distributed IBM Neffinity servers running Linux and Wndom 2000.

What Industries are using grid computing now?

Some examples include: Automotive and aerospace, for collaborative design and data-intensive testing; financial services, for running long, complex scenarios and arriving at more accurate

What are the possible benefits of a grid deployment?

Benefits can be extensive. They include:

* Accelerated time to results, which allows for the provisioning of extra time and resources to solve problems that were previously unsolvable Improved productivity and collaboration

* Allowing widely dispersed departments and businesses to create virtual organizations to share data and resources More flexible, resilient operational infrastructures

* Instantaneous access to compute and data resources to 'sense and respond' to needs using existing capital investments, which helps to ensure optimal utilization of computing capabilities

* Avoiding common pitfalls of over-provisioning and incurring excess costs

* Freeing IT organizations from the burden of administering disparate, non-integrated system

Grid and e-business on demand: what's the connection?

Grid computing is a key element in e-business on demand. Because it enables new kinds of power, flexibility and integration, IBM Grid Computing is a key element of the on demand operating environment.

Does IBM use grid computing In Its own infrastructure?

Yes. IBM is a major user of grid computing. IBM's Intragrid, based on the Globus Toolkit, is a research and development grid that allows IBM to use many worldwide assets for research purposes and help us understand the complexities of managing a grid infrastructure on an enterprise scale. Also IBM uses grids for other purposes throughout the company. One example is the IBM Boeblingen Lab Grid, composed of three IBM eServer pSeries, clusters running AIX and LoadLeveler, a cross-departmental grid used to run zSeries processor unit simulations. Jobs are submitted through a web portal, presenting users with the same interface as the one they used when running simulations on an isolated cluster. The WebSphere based portal uses the Glebus Java CoG Kit to pre-select candidate queues for submitting each simulation, using Globus Metacomputing Directory Service. This pre-selection is based on cluster loads and job characteristics. Access to a shared DB2 database allows for the automated generation of proxy certificates and for the monitoring and reporting of user jobs.

Building a grid

Building a grid can be as simple as enabling a small number of PCS (or server or storage network) to take advantage of underutilised processing and storage. This can radically speed completion of a single set of data- or compute-intensive tasks. From a relatively small deployment, you could expand slowly or quickly, narrowly or widely, depending on business needs. Ultimately, an entire enterprise can be enabled for grid--and grids can bring together not only departments and processes within a single company but also those among separate enterprises.

What about security In grid environments?

Grid Security Infrastructure (GSI) is a public-key-based security protocol, using X.509 certificates, a widely employed standard. The protocol provides single sign-on authentication, which allows a user to create a proxy credential that can authenticate with any remote service on the user's behalf, as well as communication protection and initial support for restricted delegation.

www.ibm.com

IBM