Challenges of Distributed Computing Environments - Literature review Example

Introduction To begin with distributed computing is a type of computing where computational task is divided into subtasks that are performed on a collection of networked computers. These networks under normal circumstances are used for different kinds of applications and normally they consist of Local area network, wide area networks and the internet. These networks normally differ from the other dedicated cluster interconnects (Bernard,Thalheim,JANOS,&Hans pp.203-315).A network can be described as a system where hardware and software components are networked using different computers through which allows these computers to effectively communicate and communicate and coordinate their actions only by message passing(Bernard,Thalheim,JANOS,&Hans pp.203-315) Distributed computing therefore is very much important when it comes to business deals since it tends to deliver a nonstop computing platform a business. This implies that it can be used in various situations example it can be used to scale demand by making it 100 percent available in that the loss of a few servers doesn't equate to a loss of application services. Despite the fact that work load is being partitioned across a grid of servers by installing the same application on multiple servers in a cluster, everyone is concerned with writing programs in a distributed/concurrent way to promise a scalable and highly available operation (Ackerman,pp 340-450). How It Works The computational unit of distributed computing environment is also called an entity depending on the system being modeled by the environment. Under this distributed computing entities exists which are responsible for the coordination and responses various processes and among them include a processor, a switch or an agent which are always available in the system. For the applications to effectively work well then it calls for various applications to be used among them lightweight software which is a software that is normally installed to a number of client systems. Apart from this other software’s are installed on the most used systems on the distributed computing management servers. Very lightweight software is installed on a number of client systems, and one or more dedicated distributed computing management servers. In these applications there exists requesting clients which possess some given software’s that allows the clients to submit jobs together with the required resources (Michael,Jackson,&Paul,pp.302-450). During these applications the agent dealing with processing of the application on the client is meant to quickly detect any time the system becomes idle and hence makes a notification to the management server that the system is free for the next activity to be processed and in the event an application package is requested .When this happens it implies that the client immediately receives an application package direct from the server which further runs the software when the CPU cycles becomes spare. After this action the application may run as a screen saver or just as a simple application without having any impact on the normal use of the computer (Trappl, pp.452-532). In cases where the client system is supposed to run their own applications it implies that control should be immediately returned hence the processing of the distributed application package immediately ends. This indicates that the process should be immediate since any delays in the control will actually cause unacceptance by the user. Resource identification is very much important in defining a given level of power processing, memory and given storage that each system can effectively contribute. When it comes to the issue of policy management in distributed computing environments there exists varying degrees. Administrators or other individuals who have the mandate on the job definition and how how can users access the systems can are normally ranked basing on the deadlines and perceived project value When it comes to the idea of security robust verification, encryption and sandboxing are very much important in order to prevent authorized access to data and within the distributed the systems by individuals who have no authority to do so (Yuhua,pp.16-20). Distributed Computing Technologies in Practice To start with the major technologies which have been developed comprise of CORBA, DCOM/COM, operating systems and data management systems. In really sense these technologies have been developed with a view of building distributed component based sytems.These technologies are meant to effectively allow systems to relate with other components at various levels, through provision of a given software layer and protocols that present the interoperability needed for mechanism developed in diverse programming languages hence making it easier for communication to take place effectively ((Ackerman,pp 340-450) A CORBA is the first application and usually consists of an Object Request Broker (ORB), a client and a server. An ORB is accountable for directing the requesting client to the server and making sure that the required operations takes place with reference to the object as well as its target location. Normally the ORB will examine the object position which enables it to identify target object and its remote ORB. When this happens the remote ORB invokes the method locally and sends the outcome back to the user via the network. There are many elective features that ORBs can put into practice in addition to sharing and getting remote method invocations which includes looking up objects by name, maintaining continual objects, and supporting transaction processing. The main feature that CORBA possess is its interoperability among different platforms and programming languages (Yuhua, pp.16-20). The second application is Distributed Component Object Model (DCOM) is the distributed version of Microsoft's COM technology which allows the creation and use of binary objects/components from languages other than the one they were originally written in. Normally DCOM tends to works over numerous networks through the use of different kinds of proxy’s and stubs. When the client instantiates an element whose registry access indicates that it resides outside the process space, then it puts DCOM in a position to create a binding for the element and immediately gives the wrapper and pointer to the client. The wrapper, referred to as proxy, which the marshals methods will immediately call and routes it across the network. In addition DCOM will automatically create given wrapper which is called as a stub, which unmarshals methods immediately calls and directs them to a request of the component (Ackerman,pp 340-450) A database management system is the third application an d can be referred to as a set of given computer programs which are in a position of controlling, creating, maintaining the use of computers by end users in organizations which have a given platform. This therefore indicates that DBMS helps organizations to effectively put in place control measures of the organizational wide computer databases in the hands of the database administrators and other concerned specialists.DBMS therefore is system that comprise of a software package that allows users and other software’s to store and retrieve data in a given structured way. Normally this acts as a simple way to write computer programs since it helps computer programmers to easily extract information by the user simplying asking questions in a query language. In addition provides a logic way through which organizations can freely access database information as well as using it. DBMS therefore provides adequate facilities for controlling data access through the enforcement of data integrity through concurrency management and database restoration (Ackerman,pp 340-450). DBMS therefore is a set of programs that controls the organization storage, management and data retrieval and to effectively manage them they are categorized according to the data type and structure.DBMS functions in the following ways; first DBMS accepts data requests from the application program then this program instructs the computer operating systems to take action and transfer appropriate data. Normally during the process any queries concerning the application and proper responses should be acutely submitted and received according the a given type of format which actually conforms to one or more applicable protocols .Normally DBMS allows the organization information systems to change frequently with the organizations information requirements. When the organizations information changes it provides room for new categories of data to be easily added to the database without physical making an interruption of the existing database systems and severs(Yuhua,pp.16-20). The last application is the operating system which can also be described as the interface that exists between the hardware and the user that is actually responsible for the management as well as the coordination of the other activities which actually allows the computers to effectively share resources. Normally the operating system allows for the sharing of resources with another computer which acts as a host computer. The main purpose of the host that is very important in the operating system is to handle details concerning the operation of the hardware. When this occurs it tends to relieve some application programs which have been assigned a lot o work lot hence making it easier to write applications programs without any difficulties (Yuhua,pp.16-20). Operating systems therefore is believed to offer a given number of services to various application programs and users. Normally different application of the organization will tend to use these services among them application programming interfaces and systems calls. Normally these interfaces make application request to a given service as per the operating system instructions which further pass the parameters that are intended to receive the result of the operation. Normally this allows for the users to effectively interact with the operating system and other developed software’s. For the purpose of hand held desktop computers the user interface is automatically considered to be part of the operating systems and this indicates that the large multi-user systems like Unix, and Unix-like systems are generally implemented depending on the application that the program runs which might be out of the operating system. Though most of the operating systems are currently found in cellphones and other automobiles the most common operating systems comprise of BSD, DARWIN, LINUX, SUNOS and WINDOWS NT among others(Yuhua,pp.16-20). Challenges of distributed computing environments Most challenges in distributed computing are classified into the following categories: shared memory resource allocation, distributed consensus, shared registers, computing in rings and other networks, communication protocols, and miscellaneous Other challenges of distributed computing include Addressing the impact of latency, Detecting and recovering from partial failures of networks and hosts Load balancing and service partitioning, Consistent ordering of distributed events (Yuhua,pp.16-20). Interoperating different entities in the organization among different entities, regardless of differences in hardware architectures, operating systems, communication protocols, programming languages, software interfaces, security models, and data formats has developed into one of the major issues. Making sure that the entire system appears as a single unit and the complexity and interactions between the components hidden from the end user is also another challenge (Yuhua, pp.16-20). Failure of one or more components could bring down the entire system, and therefore it should be isolated fast. Ensuring organization competence with rising number of users and accumulation of resources should increase the performance of the system. Interfaces therefore should be minimally divided and openly be available to enable uncomplicated extensions to the existing elements and other new components. By allowing the progress of tasks within a system without necessary moving the procedure of users or applications, as well as the distribution of the work load among available resources with an aim of improving performance is a challenge especially in organizations where the system is large. In addition other problems associated with distributed computing is one which is related to partial failure and the need of a fundamental reserve administrator who is in a position to ensure that adequate performance is achieved and the problems are effectively dealt with concurrently. To effectively deal with these problems it indicates that there must exist differences in the memory access paradigms between local and distributed entities which is very big problem with the fact that the distributed systems extent to ever larger sizes, hence they become more and more hard to control or even to explain (Yuhua,pp.16-20). Characteristics of distributed computing applications In a distributed computing system, a given program tends to run which is similar to other would be programs in a single computer even with the fact that different computers will tend to use different computer processors. This implies that no particular processor carries the complete load on the arrangement of resources that are meant to run on that particular program. Nevertheless, distributed computing is not truly used to reduce the work load of an individual computer's processor rather it just offers some assistance (Yuhua, pp.16-20). Most appropriate applications consist of "loosely coupled, non-sequential tasks in batch processes with a high compute-to-data ratio." The high compute to data ratio goes hand-in-hand with a high compute-to-communications ratio.This indicates that programs which have large databases that can be simply parsed for allocation are very appropriate(Yuhua,pp.16-20). It should be possible to partition the application into independent tasks or processes that can be computed concurrently. For most solutions there should not be any need for communication between the tasks except at task boundaries, though Data Synapse allows some interprocess communications. The tasks and small blocks of data should be such that they can be processed effectively on a modern PC and report results that, when combined with other PC's results, produce coherent output. This indicates that the entity tasks should be very sufficient to create a result on these systems within a few hours to a few days (Yuhua, pp.16-20). Any applications with individual errands needs given access to enormous data sets which are more proper for bigger systems than individual PCs. If terabytes of data are complex, a processor makes sense as communications can take place across the systems at a very elevated speed backplane with no bogging down the network. Server and other devoted organization clusters are more suitable for other considerably fewer data severe applications. For a distributed use using frequent PCs, the necessary data ought to fit very contentedly in the PC's reminiscence, with plenty of space to spare (Yuhua, pp.16-20). Conclusion Distributed computing systems currently are growing quickly in the last years, with main emphasis on the large development of applications and the other key impacts of internet and wireless technologies on the industry. According to Gelernter distributed computing has been slow to arrive in the mainstream computing world due to the lack of innovation to a great extent on the absence of a new computing vision from Major software vendors (Trappl, pp.452-532). Distributed systems have been vital part to peoples’ lives as an effect of innovations in the current area of Web-based applications which have continued to make a solemn force in the future. Emerging technologies therefore are expected to drive a heavy wave of improvement which in the end will enable the creation of applications that will make a bring IT as the 5th utility after water, electricity, gas, and the telephone. Distributed computing has the chance to lead into a new age in terms of computer paradigm; resource sharing pattern and online collaboration therefore identical principles and guidelines should be used, however, the specific promise of distributed computing lies mostly in harnessing the system resources that lie within the firewall. In conclusion, distributed computing is a very wide region with enormous prospective to develop competence of business processes and value of life and It will take years before the systems on the Net will be giving out compute resources as easily as they can share information (Trappl, pp.452-532). Bibliography Bernhard Thalheim, János Demetrovics, Hans-Detlef Gerhardt, MFDBS 91: 3rd Symposium on Mathematical Fundamentals of Database and Knowledge Base Systems, Rostock, Germany: proceedings Volume 495 of Lecture notes in computer science; Springer, 1991 pp 203-315 International Council for Computer Communication: Computer networks and ISDN systems, Volumes 9-10; Elsevier Science B.V., 1985; University of Michigan, Digitized11 Dec 2007.pp 102-143 Mark S. Ackerman; ACM 1996 Conference on Computer Supported Cooperative Work; Association for Computing Machinery, University of Michigan Digitized6 Nov 2007; pp340-450 Michael C. Jackson, Paul Keys, S. A. Cropper, International Federation of Operational Research Societies; Operational research and the social sciences Plenum Press, 1989 pp302-450 National Research Council, Transportation Research Record; Commission on Socio technical Systems, 1974 University of Michigan Digitized4 Dec 2007 pp. 601-610 Robert Trappl, Progress in Cybernetics and Systems Research, Volume 9 Progress in Cybernetics and Systems Research, 1982; University of Michigan Digitized15 Nov 2007; pp 452-532 Yuhua Luo, Cooperative design, visualization, and engineering: 4th international conference, CDVE 2007, Shanghai, China, September 16-20, 2007: proceedings Volume 4674 of Lecture notes in computer science LNCS sub library illustrated Springer, 2007; pp 200-321 Read More