EVENT
Technical Paper Presentation
TOPIC
Grid Computing
CONTENTS
Ø Abstract
Ø Introduction
Ø Definition
Ø Basics
Ø Importance of grid computing
Ø Benefits
Ø Security issues
Ø Optimal grid
Ø Complexity
Ø Conclusion
ABSTRACT
Grids are not futuristic? They are here now, they will work in commercial computing environments today. Based upon lack of knowledge about the convergence of grids, Web services, and self-management software, many commercial enterprises are overlooking the benefits that can be derived from grid computing”. Grid computing is an extension of distributive cluster computing. Here we focus on the basics of grid computing and places information about Grid computing into an intuitive framework, tying the pieces together, and highlighting what's important.
INTRODUCTION
We go for grid computing as many companies want to take advantage of the cost and efficiency benefits that come from a grid infrastructure today, without being locked in to a system that will not grow with their needs.
The computing grid concept had its genesis in the scientific and academic computing. This has decided implications for IT architecture. The current trend to components is supportive of the grid concept. The time frame for realization however is much longer than we would like. We can connect machines and distribute computing tasks, and distributing functionality.
GRID
The term grid, however, may mean different things to different people. To some users, a grid is any network of machines, including personal or desktop computers within an organization. To others, grids are networks that include computer clusters, clusters of clusters, or special data sources. Both of these definitions reflect a desire to take advantage of vastly powerful but inexpensive networked resources. Grid technology allows organizations to use numerous computers to solve problems by across sharing computing resources.
Current grid taps personal computers on an as-available basis to analyze data and allow the spreading of a complex calculation
The Semantic Grid is an extension of the current Grid in which information and services are given well-defined meaning, better enabling computers and people to work in cooperation. This approach is essential to achieve the full richness of the Grid vision, with a high degree of easy-to-use and seamless automation enabling flexible collaborations and computations on a global scale.
What exactly is grid computing
Grid computing is the next wave of the Internet. Think of it like a time machine. Grid computing enables you to effectively compress time by expanding computing space. Today, the first stage of grid computing -- the cluster grid -- enables organization to better utilize available compute resources. As grids evolve from cluster to enterprise to global grids -- from single departments to multiple departments to outside the firewall -- grid computing will provide seamless, transparent, secure access to IT resources such as hardware, software, scientific instruments, and services. Like electrical power, this access will be dependable, consistent, and pervasive.
Grid computing is the next wave of the Internet. Think of it like a time machine. Grid computing enables you to effectively compress time by expanding computing space. Today, the first stage of grid computing -- the cluster grid -- enables organization to better utilize available compute resources. As grids evolve from cluster to enterprise to global grids -- from single departments to multiple departments to outside the firewall -- grid computing will provide seamless, transparent, secure access to IT resources such as hardware, software, scientific instruments, and services. Like electrical power, this access will be dependable, consistent, and pervasive.
Grid computing offers a model for solving massive computational problems using large numbers of computers arranged as clusters embedded in a distributed telecommunications infrastructure. Grid computing's focus on the ability to support computation across administrative domains sets it apart from traditional distributed computing .
Grid computing involves sharing heterogeneous resources (based on different platforms, hardware/software architectures, and computer languages), located in different places belonging to different administrative domains over a network using open standards. In short, it involves virtualizing computing resources.
Today, the major focus has been on using grid computing as offloading processing capabilities across machines and taking advantage of unused capacity .The prime stated purpose for business use of the grid concept at this time is promotion of business collaboration. But a key issue is the awareness of business management
The problems to be solved might involve
· data processing,
· network bandwidth, or data storage
. The systems tied together by a grid might be in the same room, or distributed the globe; running on multiple hardware platforms; running different operating systems; and owned by different organization.
The main idea of grid computing is to grant users one place where they can go to undertake a particular task; the grid leverages its vast IT capabilities and completes the task. All the grid user experiences, essentially, is a very large virtual computer doing work.
Grid computing has the design goal of solving problems too big for any single supercomputer, whilst retaining the flexibility to work on multiple smaller problems. Thus grid computing provides a multi-user environment.
This implies the use of secure authorization techniques to allow remote users to control computing resources.
Functionally, one can classify grids as:
· Computational Grids (including CPU scavenging grids), or as:
· Data grids.
The Global Grid Forum (GGF) has the purpose of defining specifications for grid computing. Globus has implementations of the GGF-defined protocols to provide:
1. Resource management: Grid Resource Allocation & Management Protocol (GRAM)
2. Information Services: Monitoring and Discovery Service (MDS)
3. Security Services: Grid Security Infrastructure (GSI)
4. Data Movement and Management: Global Access to Secondary Storage (GASS) and GridFTP
A number of tools function along with Globus Toolkit to make grid computing a more robust platform, useful to high-performance computing communities.
| Table 1. Grid standards and toolkits | |
| | |
| | |
| | |
| | |
| | |
| Emerging Grid Applications | |
| | |
| | |
The worlds of grid computing and of web services have started to converge to offer Grid as a web service (Grid Service). The Open Grid Services Architecture (OGSA) has defined this environment, which will offer several functionalities adhering to the semantics of the Grid Service.
Grids offer a way to solve Grand Challenge problems like protein folding, drug discovery, financial modeling, earthquake simulation, climate/weather modeling etc. Grids offer a way of using the information technology resources optimally in an organization. They also provide a means for offering information technology as a utility bureau for commercial and non-commercial clients, with those clients paying only for what they use, as with electricity or water.
The Basics
Grid computing is a form of distributed computing that involves coordinating and sharing computing, application, data, storage, or network resources across dynamic and geographically dispersed organizations. Grid technologies promise to change the way organizations tackle complex computational problems. However, the vision of large scale resource sharing is not yet a reality in many areas — Grid computing is an evolving area of computing, where standards and technology are still being developed to enable this new paradigm.
Why is it important?
Time and Money. Organizations that depend on access to computational power to advance their business objectives often sacrifice or scale back new projects, design ideas, or innovations due to sheer lack of computational bandwidth. Project demands simply outstrip computational power, even if an organization has significant investments in dedicated computing resources.
Even given the potential financial rewards from additional computational access, many enterprises struggle to balance the need for additional computing resources with the need to control costs. Upgrading and purchasing new hardware is a costly proposition, and with the rate of technology obsolescence, it is eventually a losing one. By better utilizing and distributing existing compute resources, Grid computing will help alleviate this problem.
Benefits of Grid Computing
· Better utilize computer resources (including current assets and future purchases)
· Increase efficiency
· Solve problems more quickly
· Get products to market faster
· Increase access to resources
· Better control over resource allocation
· Allow for more efficient deployment across heterogeneous environments
· Enables joint collaboration projects in a dynamic, distributed, virtual organization.
Security Issues
Security remains an issue for the industry to better address. Most chief information officers, do not want to expose their corporate assets over the Internet if it isn't secure. In trusted environments, within a department or within an open research community for example, where applications and data are not in a mission-critical or proprietary state, security is less of an issue. Beyond those safe harbors, there is a call for strong security measures -- and the industry is working to overcome both real and psychological firewalls. Currently, several working groups within the Global Grid Forum are working on standardization of the many already existing security solutions.
Grid computing harnesses a diverse array of machines and other resources to rapidly process and solve problems beyond an organization’s available capacity. Academic and government researchers have used it for several years to solve large-scale problems, and the private sector is increasingly adopting the technology to create innovative products and services, reduce time to market, and enhance business processes.
Although the computational problems solved today by grid computing are often highly sophisticated, the software available to manage these problems cannot handle connected parallel applications. As it turns out, creating a parallel application to run on a grid is even more difficult than creating a large monolithic custom application for a dedicated supercomputer or computer cluster.
Optimal Grid
Most people working on grid computing today focus on the challenges of its physical operations, such as how to determine what computer and database resources are available and how to organize them into a functioning system.
Most people working on grid computing today focus on the challenges of its physical operations, such as how to determine what computer and database resources are available and how to organize them into a functioning system.
 | |
| Figure 1. Optimal Grid has been used to model the propagation of infrared light through a photonic-bandgap structure of silicon pillars (represented by circles), a calculation that requires interactions between electrical and magnetic fields of nearest-neighbor grid cells and which grows rapidly in memory demands and run-time. The peaks show the value of the magnetic field, which is related to the intensity of light. (Geoffrey W. Burr, IBM Almaden Research Center) | |
To demonstrate how one might simplify the creation of applications on a grid, prototype called Optimal- Grid has been developed, which handles both independently parallel and connected parallel problems (Figure 1). Optimal Grid.. Optimal-Grid requires only that the networked computers all have a Java run-time installed.
Not even an expert administrator could orchestrate the complex connected problems of a heterogeneous distributed-computer system. To this end, the Optimal Grid system incorporates instrumentation, feedback, and a certain amount of knowledge, or rules, to maintain a balanced performance on the grid and react to various kinds of failures
Users simply have to supply the code that represents their basic problem algorithm, and Optimal Grid manages everything else. Each node on the grid receives a piece of the problem, which consists of a collection of original problem cells (OPCs) (Figure 2). An OPC is the smallest piece into which the problem is divided, and each one needs to communicate and share data with its neighbors. Optimal Grid automates this communication and attempts to minimize the amount of network communication needed to solve a problem.
. When the Optimal Grid system initializes itself to solve a problem, it automatically retrieves from the grid a list of available computer nodes. It also obtains the grid’s performance characteristics. Optimal Grid allows the system to self-heal if one or more computer nodes fail during a computational sequence. When Optimal Grid detects the failure of a computer node, it stops calculations across the grid until the failed node recalculates the results lost during the sequence. Although the grid must remain idle during this catchup phase, a short delay is preferable to having to restart the problem solution from the beginning. Once the node finishes its recalculation, the grid continues working on the overall problem.
 |
| Figure 2. A set of methods describes the connectivity of the original problem cell (OPC) with its neighbors and specifies the calculations to be performed by the cell using local data. Groups of OPCs form collections, one or more of which define the variable problem partition assigned to a computer node. |
The Optimal Grid system is designed to bring the immense potential of grid computing within easy reach of users who are not grid-infrastructure experts.
By including autonomic features such as self-configuration, self-optimization, and self-healing, OptimalGrid seeks to deliver a robust system capable of handling truly connected problems to meet a broad class of user needs for a broad range of industrial and scientific applications. OptimalGrid is a new programming model designed for the grid environment. It is optimal in the sense that the system attempts to optimize and balance the pieces of the workload to make the best use of any existing grid infrastructure. Initial results look promising.
Complexity in Grids
The simplest class of applications addressed with a computational grid has been independently parallel problems. These applications work in a simple scatter–gather model; that is, a problem is divided into pieces of data, and a separate data set is sent via the Internet to different nodes, each of which works independently and without communicating with the other nodes to derive its results.
The simplest class of applications addressed with a computational grid has been independently parallel problems. These applications work in a simple scatter–gather model; that is, a problem is divided into pieces of data, and a separate data set is sent via the Internet to different nodes, each of which works independently and without communicating with the other nodes to derive its results.
Such problems are well suited for the distributed computing power of a grid. A larger and more general class of applications can be described as connected parallel problems, which require more sophisticated management in almost every area, including problem definition, problem partitioning, code deployment, grid–node management, and system coordination.
CONCLUSION
Grid computing is becoming a critical component of science, business, and industry. Making grids easy to use could lead to advances in fields ranging from industrial design to systems biology to financial management.
Grids could allow the analysis of huge investment portfolios in minutes instead of hours, significantly accelerate drug development, and reduce design times and defects. With computing cycles plentiful and inexpensive, practical grid computing would open the door to new models for compute utilities, a service similar to an electric utility in which a user buys computing time on-demand from a provider.
Some industrial applications are important enough to warrant the use of dedicated high-end computers (supercomputers or clusters of computers and/or supercomputers). A much larger body of scientific and engineering applications stands to benefit from grid computing, including weather forecasting, financial and mechanical modeling, immunology, circuit simulation, aircraft design, fluid mechanics, and almost any problem that is mathematically equivalent to a flow.
.
.
.
No comments:
Post a Comment