Grid Topologies - Best Essay Writing Service Reviews Reviews | Get Coupon Or Discount 2016
Free Essays All Companies All Writing Services

Grid Topologies

The concept of Grid computing is very generic in nature. Some organizations are revising their own grid system so that it fits in their needs. In an organization, a number of computers are connected to one another in order to share resources. The intra-grid topology is configured in a single organization. The computers are connected through a high-speed locally configured network. In this topology, the design of a scheduling system is much easier since it provides static computing resources between computer systems.

An intra-grid is a single administrative domain and the locally configured network which is designed privately guarantees a bandwidth. (http://oslab. khu. ac. kr/mgrid/resources/rao_ZHU_Yanmin_Survey_Report. pdf) The main idea of this is a unity among servers, networks and some storage systems into a single one for as faster delivery of information. The complicated system of an organization can be transformed into a single virtual computer which can help solve complicated problems. (http://www. ibm. com/developerworks/grid/newto/) The idea of grid computing is to allow computers to work together.

The virtualization of computers allows an organization to enhance development of older system and to enable the development of a new one and utilize the resources which are available to the users. The administrators can benefit from this since this helps provide a resilient infrastructure and it also improves the access to the data and information and it supports collaboration across different disciplines. The idea of this also allows business establishments to increase the use of their existing resources. (http://www. ibm. com/developerworks/grid/newto/) 4. 2 Extra-grid

It is always the goal of grid computing to implement systems which can be help to everybody who is a part of the organization. With this, grid computing keeps on grouping computers and keeps on conducting researches in order to find ways on how to make things better and to improve transaction processing. If two or more intra-grids are configured, an extra-grid exists. If an intra-grid is designed a single administrative domain, an extra-grid is designed for more than one administrative domain which makes the level of management issues more complicated as well as its complexities increases.

The resources in this topology become more dynamic in nature. An extra-grid involves a dispersed security, multiple domains and Wireless Area Network (WAN) connectivity. (http://oslab. khu. ac. kr/mgrid/resources/rao_ZHU_Yanmin_Survey_Report. pdf) The principle of grid computing involves networking file systems to provide ease of information dissemination, increase communication and bandwidth and increase the processing power of all systems used in the organization.

This also may be an instrument to decrease the cost in the development of new applications and increase the speed of application so that the response becomes faster as well. In addition, this also allows the management for a high utilization of the company’s existing resources. (http://www. ibm. com/developerworks/grid/newto/) The principle of an extra-grid always has an impact in the society since the result is very evident. 4. 3 Inter-grid Another topology adopted by the grid computing is the inter-grid which has the same concept as the internet.

Since it has the most complex form of topology, it has dispersed security, multiple domains as well as a Wireless Area Network (WAN) connectivity. If in an organization, there is a need for collaboration among the various computing community, the establishment of an Inter-grid is necessary. This can also simplify and end-to-end process within the organization. (http://oslab. khu. ac. kr/mgrid/resources/rao_ZHU_Yanmin_Survey_Report. pdf) The internet offers a variety of information and allows users to share ideas and files. The concept of inter-grid is somewhat the same as this.

This allows the computers to communicate in a much easier and faster way. Grid computing allows the management to provide a capacity for application which demands are high. This, additionally, helps the organization in the creation of a virtual one so that data and other resources can easily be shared. In some organizations, grid computing can speed up transactions as well as the provision of stable IT environment which can be used to provide a mission-critical environment. (http://www. ibm. com/developerworks/grid/newto/)

The figure below shows the relationship between these three topologies: Figure 1: Relationship among Inter-grid, Extra-grid and Intra-grid Topology 5. Grid Computing Application Considerations: 5. 1. CPU considerations The implementation of this kind of technology requires a much faster processor since multiple systems need to run simultaneously. The advancement in operating systems, processor technology and compilers are capable of providing a multiple processing benefits, thus lessening the burden on the part of the programmers who do the application.

Some distributed systems are theoretical in nature. The symmetries in Base One can reduce the complexities it brings, in the architecture itself. The various aspects of symmetry translate into codes with considerations in equivalence of batch jobs servers, back-end Database Management System and a unified client/server environment. (http://www. boic. com/b1mgrid. htm) Another feature of the said architecture is the role of a database in virtual supercomputers which simplifies work synchronization of multiple processors.

It also controls the power of a DBMS technology in which the reliability and security are properly addressed. (http://www. boic. com/b1mgrid. htm) It takes a high speed processor to make this system run. The system is capable of running multiple and simultaneous programs. 5. 2. Data considerations The objectives of the design includes a local implementation of a DBMS in order for a single application to operate against databases which are purchased from different vendors and it is expected to provide and increase the performance rate of caching without the need for placing the burden to the programmers.

The network communication is handled properly in a data-centered model which works within the core of the databases. (http://www. boic. com/b1mgrid. htm) Additionally, the components of a grid computing include security, user interface, workload management, scheduler, data management and resource management. The data management facility of the system handles secure and reliable data which moves from one right place to another. Moreover, to locate computers to run an application and to assign a job, a scheduler is needed.

The task of this is just to assign the next resource to be allocated in the system. Some other essential services include bandwidth allocation, processor request and data and storage management. These are the services offered by grid computing which adds functionality in web services. (http://www. boic. com/b1mgrid. htm) 6. Grid Computing compared to 6. 1. Distributed Computing Grid computing is basically designed to utilize CPU cycles of computer systems. In the early 1980s, CPU power and bandwidth makes distributed computing a practical idea.

In distributed computing, thousands of computers are managed and pooled which have a limitation in their memory and power. Moreover, distributed computing is where a group of interdependent computers are linked to solve certain complex problems. Furthermore, it can be termed as a system which provides computers which collaborate over the network and a system which involves various architectures that share data resources. (http://www. jatit. org/distributed-computing/grid-vs-distributed. htm) Distributed computing is where the utilization of an idle CPU and storage space are networked to work together.

Due to a lack of application, the growth has been limited to some extent. On the other hand, the increase in CPU power and bandwidth has made distributed computing a more practical idea. (http://www. extremetech. com/article2/0,2845,1154107,00. asp) There are various types of distributed computing applications. One scenario which describes one type is that a search in the database can be split up across different desktops which can then be submitted concurrently against a fragment on every desktop computer.

In addition, an exhaustive search technique which requires searching through a huge database application also describes a type of distributed computing. A complex financial modeling also is an example of such. (http://www. extremetech. com/article2/0, 2845, 1154107, 00. asp) There are some applications, though, which are not suitable for a distributed computing. When an application is near to its running stage, the less appropriate it is to be considered as a distributed computing.

The characteristics of an application which can belong to this type are the one with a “loosely-coupled, non-sequential tasks in batch processes with a high computer-to-data ratio”[ (http://www. extremetech. com/article2/0,2845,1154107,00. asp)]. The latter goes with a high compute-to-communications ratio. In addition, the application should also have the ability to take advantage of a coarse-grained parallelism which allows the application to be partitioned into smaller and more independent tasks which can easily be processed simultaneously.

The data must be small enough to be processed on a modern computer. (http://www. extremetech. com/article2/0, 2845, 1154107,00. asp) Figure 2. Distributed System vs. Peer-to-Peer and Client-Server Computing Figure 3. Distributed Computing over the Web (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) 6. 2. Cluster Computing If grid computing focuses on its ability to support computation across various domains, computational grid is focused on operations which require an intensive computation. A cluster computing only happens when two computers are used to solve a problem.

Cluster computing is also heterogeneous. Moreover, in cluster computing, all computes must have the same hardware and Operating System (OS) running, unlike in grid wherein each computer may have different hardware and OS running. Cluster computing is dedicated to work on a single machine only. Some characteristics of a clustered computing include a tightly coupled system, a single system image and a centralized job management scheduling system. (http://www. jatit. org/research/introduction_grid_computing. htm) 6. 3 Peer-to-Peer Computing

A peer-to-peer refers to systems and applications which are employed a distributed resource in order to perform functions in a decentralized manner. A peer-to-peer computing enables valuable externalities which work through the aggregations of resources through a low-cost interoperability, lower cost of ownership and sharing of cost which uses the existing infrastructure, thus eliminating the maintenance costs. And most importantly, this promotes anonymity and privacy through some algorithms and designs which allows a degree of autonomous control over data.

A peer also gives out resources and at the same time, gets resources from other sources. This is also another way of computing, which can treated differently when compared to a centralized and a client-server model. It is further illustrated below. (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) Figure 4. Simplified Peer-to-Peer and Client-Server Computing (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) In addition, peer-to-peer is also concerned on the general evolution of computing in which computing is done at the edge of internet.

Figure 5. Peer-to-Peer Solutions (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) The peer-to-peer computing also includes some technology requirements which include a scale of deployed computers, an ad-hoc connectivity and security, architectural requirements which include the availability of computers, scalability of future systems and privacy and anonymity and economy requirements which includes the cost of ownership and pervasive use of the system. (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf)

The Peer-to-Peer model can be classified into either pure or hybrid. The former model does not need to run a centralized server while the latter, a server is needed to obtain information. The figure below shows taxonomy of a Peer-to-Peer system. (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) Figure 6. Peer-to-Peer Computing System The architecture of a peer-to-peer computing is depicted in the next figure. (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) Figure 7. Peer-to-Peer Architecture (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1.

pdf) Figure 8. Comparison of Solutions (http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf) 7. Future Trends Technology keeps on advancing as years go by. All these are aimed at providing the best services and products to all users. All these are designed to make transactions faster and easier. Management keeps on innovating strategies in order to cater the needs of their clients. With this, the future trend of grid computing has greater chances that clients will love its implementation. Many people have been applying the concept of grid computing.

Grid computing, in the future, will tend to have an impact in Enterprise Resource Planning (ERP) or Customer Relationship Management (CRM). This will then enable newer solutions to problems based on outsourcing solutions between providers of different applications. Furthermore, Foster [CIO News, 2003] quoted the following: (http://searchcio. techtarget. com/news/interview/0,289202,sid182_gci895039,00. html) ? “Increasingly, people will work out how to build various forms of service provider offerings. No one has been able to do that effectively to this point. ” ? “Starting internally on a small level is the most promising direction. “

The intersection of grid computing technology and web services is with the implementation of an Open Grid Services Architectural (OGSA) which will then integrate grid technologies with the web services. As years go by, web service is now out in the open and which has a big potential in a distributed environment. An enterprise must be able to connect the computers in one place only. One of the most successful organizations such as IBM, HP, Sun and Oracle are into the innovation of grid computing strategies and applications which can surely increase productivity. (http://searchcio. techtarget. com/news/interview/0,289202,sid182_gci895039,00.

html) With the benefits and advantages of grid computing, Sun Microsystems are offering the best services when one makes a purchase of computer. This is because of the advances offered by grid computing technology. After the concept of grid computing has been implemented, there was a tremendous growth and a more efficient way of handling transactions. For IBM, they have a product which is called the IBM grid toolbox which has some enhancements regarding IBM. Because the disadvantages outnumbered the advantages of grid computing, more and more users are into the implementation of such technology.

Data management will eventually become easier to handle as well as the transfer of information between one computer sources to another. More and more companies are adapting the idea of grid computing. Computer problems can then be easily solved. References: Erlanger, Leon. 2002. Distributed Computing: An Introduction. Retrieved on November 29, 2008 at http://www. extremetech. com/article2/0,2845,11769,00. asp. Hines, Matt. 2003. Expert Charts future for Grid Computing. Retrieved on November 27, 2008 at http://searchcio. techtarget. com/news/interview/0,289202,sid182_gci895039,00.

html. Milojicic, Dejan, et al. 2003. Peer-to-Peer Computing. Retrieved on November 28, 2008 at http://www. hpl. hp. com/techreports/2002/HPL-2002-57R1. pdf Worthington, David. 2005. Interview: The Future in Grid Computing. Retrieved on November 28, 2008 at http://www. betanews. com/article/Interview_The_Future_in_Grid_Computing/1109004118. Yanmin, ZHU. No Year. A Survey on Grid Scheduling System. Retrieved on November 28, 2008 at http://oslab. khu. ac. kr/mgrid/resources/rao_ZHU_Yanmin_Survey_Report. pdf. No Author. No Year. Comparison of Grid Computing vs.

Cluster Computing. Retrieved on November 29, 2008 at http://www. jatit. org/research/introduction_grid_computing. htm. No Author. No Year. Difference between grid computing and distributed computing. Retrieved on November 29, 2008 at http://www. jatit. org/distributed-computing/grid-vs-distributed. htm. No Author. No year. The Base One Grid Computing Architecture. Retrieved on November 28, 2008 at http://www. boic. com/b1mgrid. htm. No Author. No Year. Introduction to Grid Computing. Retrieved on November 28, 2008 at http://www. ibm. com/developerworks/grid/newto/.

Sample Essay of