Abstract:
Grid Computing is a term referring to the combination of computer resources from multiple administrative domains to reach a common goal. The grid can be thought of as distributed systems with non-interactive workloads that involve a large number of files"[1]. In the olden days supercomputers were used to do many complicated tasks. This process is called as parallel processing where many CPUs involve. Grid computing can be called as a special type of parallel processing which depends on complete computers which are connected to a network by a conventional network interface[1].
Introduction:
The growth of the countries all over the world made the growth of the internet along with the availability of many resources like computers, and high speed internets for very low price. This made the engineers to think about new way of changing the information from one place to other and one country to another. For this they have combined or clustered a wide variety of geographically distributed resources which includes supercomputers, databases, storage devices and many other instruments which can then be accessed as a single resource.[2]. This new way of paradigm is termed as GRID computing. Grid computing is mainly uses the middleware to divide and assign pieces of tasks and programs among several computers. Grid computing is a different form of distributed where a super virtual computer is connected to many different computers performing on same tasks.[2] It is also believed that the next generation of internet would be based on Grid technology. It is determined to solve many problems related to information sharing and collaboration between the resources. [9]
Grid Middleware:
Grid Distributed Middleware:[3]
Grid will provide high level resource sharing ability. Globus Tools will support as interface for some basic applications of Grid but it is difficult for development of complicated grid applications. Computing grid distributed middleware (CGDM) was later proposed by some of the engineers which will act as the bridge between the computing grid application and Globustool kit.[3] Architecture and mechanisms if CGDM will help the Grid to perform at best level. The kernel of CGDM includes Container, Plug-in, Object pool, Grid Object adapter and Distributed middleware Protocol. [3] The Architecture of the CGDM consists of Hiberarchy of CGDM which again subdivided into CGDMP layer, Support Module layer and container and component layer. The following figure shows how it looks
Figure 1 The Hiberarchy of CGDM [3]
With the help of CGDM portal will only make all the process operations and the data transfer with the end users.
Light weight Grid middleware Based on OPENSSH-SCE: [4]
This is another type of grid middleware which is based on Scientific Computing environment (SCE) for the scientific supercomputing service environment.[4] This mainly works on OPENSSH package and will adopt the open grid services architecture. This middleware mainly depends on service based middleware.[4]
Figure 2: SCE Architecture.[4]
The SCE Architecture has three layers which are Interface, Center server and SCE Nodes. The Interface will acts as user interface where browser will act as medium. The center server mainly provides the four services like Grid Security, Information service, Job management, and Data Management. Among all Job Management is core part of this middleware. This SCE provides more security when compare to other Middleware's .This SCE was under research scientists have just implemented on Information service and Job management. Whenever user wants to use this middleware for performing a job he need to login and find the grid service which he needs to perform. This middleware is under development process and will be used in many applications if it is completed.[4]
A CORBA based GRID Middleware Architecture: [5]
This is a Grid middleware which assures the quality of services requirements such as availability, reliability etc. To give support for the future grid applications there is a need for the development for new grid middleware that will improve the level of abstraction will reduces the complexities of the previous middleware's. This middleware mainly uses the standards based CORBA distributed object computing and integration technology to design a new middleware GRID TAO (GRIDT). This middleware support many future challenged Grid applications.
Figure 3: Grid Middleware TAO [5]
This middleware is designed to improve and develop low level Grid middleware such as Globus tool. The main idea behind this GRIDT is to provide GSP and MRB. The GSP offers a web service interface for all grid applications and the "meta resource broker (MRB) manages virtual resources that are abstractions of concrete resources belonging to multiple service providers".[5]
Types of Grids:
There are many types of grids. There is no particular standard way of categorizing the grids. Different scientists, organizations have divided these grids according to their application requirements.[6] These grids are not strictly divided they can be in two or more forms.
Computing grids: These grids are used for the computational purpose of the CPUs. These Computing Grids will reduce the Computations for many organizations. These Computing grids are used by engineers, scientific and space research organizations [6].
Information Grids: These grids are used for collaborative computing file sharing. These are also known as data Grids[6]
Service Grids: These Grids are used for combining physical elements of Grid connection to provide same operating system for different systems across the world.[6]
Intelligent Grids: These are the grids which will consist of basic grid network which automatically manages the network.[6]
Cluster Grids: These are the grids which will consist of one or systems which are working together on the same task which also provides the same point of access to users. These cluster grids support high performance and throughput. [6]
Global Grids: These are the grids which can be implemented in most of the applications .It provide the resources such that we can implement anywhere in the world.[6]
Applications:
For the grid computation to be implemented the tasks should be divided into sub processes and run in parallel. Grid Computing is followed and implemented in many major technologies, projects all over the world. It has many applications. The following are some of the applications of Grid Computing. There are lots of applications where grid Computing is implemented.[7]
Human Genome Project (HGP):
This is one of the applications where grid computing is applied. "The human genome is composed of 24 distinct chromosomes with about 3 billion DNA base pairs organized into 20,000 - 25,000 genes"[7]. In this project they have implemented Grid Computing to identify the sequence of genes. They have completed three years ahead of the target period.[7]
In chemistry:
In chemistry they have been doing researches on molecular structure, molecular behavior and different reactions taking place. It takes many years if that research is done at one place. They are implementing grid computation for this which will save their time and also money. Computational grid is one of the organizations which will provide all the materials and resources for computational Grid [7].
In Business:
Grid computation plays major role in Business applications where the organizations will gather their processes, data which will be processed in parallel. Data management plays a major role in any business. Oracle has best database management system in the world which is used by many organizations. Oracles 10g is very famous in which 'g' stands for grid which is the one of the management system for grid computation [7]. Grid computation works in virtual environment where all the resources can be gathered and implemented.[7]
Pro-Active Mobile Computing:
Grid computing has many service oriented architectures. It provides accessibility of vast computing and resources all over the world. This feature made Grid computing to support for wireless mobile applications. All users can access the grid from their mobile devices when they are away from company, organization. This is very effective in business applications where data can be retrieved from repositories [8]. Mobile Grid access was developed to provide a collaborative problem solving environment which is the main idea of grid computing [8]. Let's see one of the applications of this Mobile Grid Computing
They developed a Grid based mobile computing demonstrator which is named as Distributed Aircraft Maintenance Environment (DAME) project. This DAME mainly does diagnosis and prognosis problem. This mainly proved that Grid enabled computing can be very effective for many organizations, business applications by providing access to many powerful grid services under different situations.
Figure 4: Distribution of data in a virtual aircraft maintenance environment [8]
Geographical Information service:
Gird computing mainly focuses on the collaborative work which is mainly used in high performance computing area. Grid Computing has great features which supports in building a Geographic Information Grid Service (GISG). This has four layers which has its own functions to be done. Geographic information has very large data and also high computing methods which are very difficult to share information through web. They found that Grid technology has great potential to overcome the difficulties in sharing large information around different places in the world. [9] "From the point of view of National Spatial Information Infrastructure (NSII), the study on the Geographic Information Service Grid is a feasible direction to solve the geographic information sharing and service based on the strong Grid technology".[9]
Figure 5: Four layer model to construct GISG [9]
Security:
Grid computing is widely used in different fields. In grid computing, geographically distributed of same resources are computed as a whole. Due to its wide range of applications enterprises are worried about its security issues. Grid computing at first was used by academic purposes later was moved even to enterprise communities [10]. Computer system which usually includes the PCs and servers is known as Host." Grid resource is a computing or data element within the grid. Credentials are used tickets or tokens used to identify, authorize, or authenticate a user" [10] . The Grid security is broadly classified into host, architecture, and credential levels.
Host level:
It is related to the on-demand grid system in which the host will have contact with grid. This host level has two types of jobs local and remote jobs. Security issues arise between these two while sharing of memory taking place such as local data corruption
Architecture level:
This security is concerned about the security issues related to architecture such as mapping policies, denial-of-service within the grid system, and also like information issues like reliability, integrity, and authentication. [10]
Credential level:
"Credential-level issues become very important in a grid context because multiple systems require varied credentials for users to access them". [10]
Grid computing security is main issue which should be a goal for any application which is implementing Grid computing.
Figure 6 Taxonomy of grid computing security issues[10]
Precautions:
The organization should be able to identify specific vulnerabilities, threats, and policy issues in their grid implementation.
Should develop some countermeasures based on different threats.
Should be evaluated in regular intervals.
Should follow some rules in implementing Grid computation.
Conclusion:
Grid Computing is one of the technologies which is getting popular day-by-day. It has been implementing in various fields like science, Business and Research purposes. As there are various types of grids organizations can use them according to their applications which is very flexible feature. Grid computing is very important in present scenario where it has very strong features which support data sharing, data management and many other which makes a difference for all applications where it is used.
