The distributed computing technology are widely been used by big companies all over the world. It has been one of necessary things required to help the users to be always competitive to their competitors. One of the new technologies used today is Grid Computing. By using the grid computing, the user can take control on computing resources such as processors, storage, data, applications, and others as needed without any knowledge about locations of those resources or what the hardware used, operating system , and so on.
Grid computing uses middleware to coordinate various IT resources all over the network, allowing them to function in virtual. The resources used are function as single system in front of users. There were a Telescience Project that provides remote access to very powerful electron microscope at the National Center of Microscopy and Imaging Research in San Diego. The Grid's users can operate the microscope, allowing new access to the instrument and its capabilities in remote. This is one of examples about grid computing.
An environment that provides the ability to share and access resources across a distributed and heterogeneous environment transparently, requires the technology to virtualized certain resources, technologies and standards in the areas of security, scheduling , systems management, accounting, and others.
Benefits From Using Grid Computing
Many big companies such as Dell Inc., Intel, Compuware Corporation, Oracle Corporation, and others are using grid computing. It is because there are many advantages can be derived by using the technology. Instead of using one heavily burdened server, the load can be spread evenly across many smaller computers to be processed. Many advantages can be derived from types of services provided by Grid such as:
Grid provides Computational services by providing secure services to execute application tasks on distributed computational resources individually or collectively. Resources brokers provide the services to allow the distributed resources to be used collectively. These types of Grid are also called a Computational Grid.
Application services. The services keep the transparency by hiding the provided access to remote software and libraries from users. The application services are defined by Web services that are always growing day by day. An example system that can be used to develop such services is NetSolve.
Data services. The data services are concerned by proving security of access to distributed datasets and their management. The data will be replicated, catalogued, and different datasets are stored in different locations to increase the availability of data access, and security of data.
Information services. The data are extracted and presented by using the services of computational, data, and/or application services. This is the way how the information is represented, stored, accessed shared, and maintained.
Knowledge services. These services are related to how the knowledge is acquired, used, retrieved, published, and maintained to help users in achieving their goals and objectives. Knowledge is understood to solve a problem, achieve goals or execute a decision. Data mining is a good example which is responsible in building a new knowledge automatically.
Figure 2-1 Grid Computing
Types of Resources
A grid is a collection of devices, sometimes referred to as nodes, resources, and many other terms. They all are connected together to form a single Grid system. Some resources can be used by all grid users and some of them may have specific restrictions. The resources of Grid Computing can be divided into few types of resources.
Computation
The most important thing that gives a big influence on the performance of grid computing is processors on machine in the grid. The computing cycles provided by those processors affects the grids processing power in tasks executions.
Storage
Another resource that also has a big influence on Grid performance is data storage. There's a term related to data storage called data grid. The data grid is an integrated view of data storage provided by a grid. Each machine exists in the grid usually provides some storage quantity to be used by grid even if it is temporary storage. The storage used can be consisting of secondary storage, or other permanent storage media or it can be a volatile such as a memory attached to the processor. The volatile memory can be used to cache data for running applications. While the secondary storage plays the major role to enhance the performance, increasing the capacity, sharing, and reliability of data.
Communications
Data communication capacity is one of important resources of a grid including communications within the grid and external to the grid. The available bandwidth is a critical resource that can limit the grid utilization. The data movements are depends on the bandwidth. If the bandwidth available are in small capacity but the amount of data to be processed are large, then the data flow's continuity may be affected.
Software and licenses
The cost can be too expensive if the software required is installed on each machine in grid. By using grid, only a particular of grid machine have the software installed on it. The tasks that require this software to be processed are sent to the machines which contains the software. This approach can save the cost of licensing fees.
Special equipments
The grid system uses special equipments, operating system, and different architectures that represent different kind of resources that will be used to assign tasks to the machine. Some software may be available only on a particular type of machine. The attributes must be considered before assigning jobs to resources in the grid. Sometimes, the grid administrator may develop a new artificial resource used to assign tasks according to policy rules and other constraints.
Figure 3-1 Grid Application
Grid Construction
There are some main aspects need to be considered during Grid construction. Those aspects will determine the grid characteristics. Those aspects are stated as follows.
Multiple administrative domains and autonomy
The grid resources are geographically located at different locations and connected through multiple administrative domains and owned by different organizations. The rights of resources owners must be honored with their usage policies and local resource management.
Scalability
The grid may grow geographically as addition of more resources and facilities in different locations to enhance the performance of the system.
Heterogeneity
Resources used in grid system are consists of various types of devices. These entities must be able to communicate to one another, despite of differences in hardware architectures, communication protocols, operating systems, data formats, and others.
Fault tolerance and failure management
The system should be able to face any possible failures. Any occurred failures in some part of the system should not able to bring the entire system down.
Security
Security is always a major issue when considering a grid environment's security. The common security components used are mechanism for authentication, authorization, and confidentiality of communication between components in the grid.
Design
Grid system usually is built based on the existing components. A few aspects need to be considered during design phase are the grid architecture models, grid topologies, and others.
Grid Architecture Models
The grid architecture models can be spread into two types. Each type has its own design purpose. Some grids are designed to take advantage of extra processing resources while the others were design to support cooperation between different organizations.
Computational Grid
These types of grid gather the processing power from the distributed systems. SETI@home grid is an example for computational grid. The system consist of low-powered computer systems with minimal application logic and small storage capacity. Many business companies are interested in computational grid due to its ability to maximize the computer utilization of existing resources.
5.1.2 Data Grid
The data grid system is expertise in providing secure access to heterogeneous pools of data. The data grids also can contain resources like federated database. Other than that, the data grid also can maximize the usage of data, storage, and network resources located in distinct administrative domains, and provide a quick and reliable data access. Data grid system have a big potential to be expanded on its data-mining abilities and maximizing the utilization of existing storage devices and to reduce the complexity of data management.
Grid Topologies
Grid topologies can be expanded into three types as stated in Figure 5-1. There are :
Intragrids
Extragrids
Intergrids
Intergrid
Extragrid
Intragrid
Figure 5-1 Grid Topologies
5.2.1 Intragrids
Intragrids used by single organizations. All of nodes or computers included in this grid are using the same security domain, and share data internally on a private network. The bandwidth on the private network is high with high availability. Its just built on single cluster of machines.
Figure 5-2 An Intragrid
5.2.2 Extragrids
Extragrids contains more than one intragrids. Extragrids consist of multiple cousters and involved more than one security provider. The level of management complexity are higher than intragrid. It is different from intragrid, extragrid used by more than one organizations and have remote or WAN connectivity.
Figure 5-3 An Extragrids
Intergrid
Integrids used by many organizations with multiple partners and multiple clusters. The security is dispersed all over the grid. The data contained is a global public data which can be accessed by all authorized users.
Figure 5-4 An Intergrids
