Research into the Software Required to Support a True Distributed Computer System - Essay Example

?Research into the Software required to support a True Distributed Computer System By Table of Contents Research into the Software requiredto support a True Distributed Computer System 1 Table of Contents 2 Distributed Computing Environment: An overview 4 Utilities of Distributed Computing Support SoftWare 5 Layered Structure of Distributed Computing Software 7 Software Architecture 9 Distributed Computing Application Characteristics 10 Distributed Computing Software Types 12 Distributed Computing Operating Systems 13 Distributed Computing Services Management Software 13 Distributed Computing Computational, Storage, and Networking Resources management Software 14 Distributed Computing Application Development Software 15 Distributed Computing Security Management Software 15 Distributed Software Data Movement Software 16 Distributed Computing Resources Management Software 16 Distributed Computing Remote Access Software 17 Distributed Computing Multimedia Access Management Toolkit 17 Conclusion 18 Ahmad, I. and Lau, F., 1999. Special issue on software support for distributed computing. Journal of Parallel and Distributed Computing - Special issue on software support for distributed computing, 59(2), pp.101-06. 19 Introduction Ever-increasing competition and quick developments in information and communication systems, technologies, and a great deal of proliferation of low-cost workstations and PCs have offered a broad avenue for distributed computing to develop into a conventional computing structure and arrangement. At present, distributed computing is no longer limited to literature and advanced research laboratories; however, it is now easily in reach of a huge community of technology producers and clients. In fact, the transformation of status would not have been possible if there were not considerable endeavours that technology and communication researchers have offered for the development of essential supporting application, distributed OS, remote procedure calls, network file systems, OO (object oriented) middleware and message passing interfaces (Ahmad and Lau, 1999). In addition, the arrival of the WWW (World Wide Web) and high speed broadband Internet connections have as well put fuel to such kind of technology transformation. Presently, the Internet has turned out to be one of the biggest distributed systems. Currently, it is being utilized by the majority of people to resolve very complicated and challenging issues such as “search” to discover the biggest prime number, intended for space intelligence, and to break the RC5 block encrypted cipher, etc. In this scenario, a lot of huge and worldwide distributed systems are in progress to develop a huge “Globe system.” This Global system will offer an excellent support for resolving global issues (Ahmad and Lau, 1999). This essay discusses the software required to support a true distributed computer system. This research will provide a detailed overview of true distributed computing environment that is a necessity to develop and use appropriate co-operating subsystems. Distributed Computing Environment: An overview The distributed environment is in some sense overcoming the limitations of traditional computing offering independence from the Personal Computers (PCs) to the golden or dark period of the mainframe. In the past, computers were single user machines with minimum storage capacities. Hence, the appearance of time-sharing machines and operating systems provided excellent opportunities to people all around the world. In fact, it allowed them to make use of the same hardware framework to support several people working simultaneously or cooperating with each other. Additionally, with the arrival of wide area networks (WANs), this cooperation further extended across the world and up till now simply needs a few central points of integrity control and management (Trireme, 2004; Nemeth and Sunderam, 2002). In addition, the arrival of the workstations took the people right back to single user machines where they were able to take their backups and so on. Therefore, they were once more isolated from other users with whom they might want to communicate and share. Though, the arrival of the web and LANs helped resolve these issues, however, it brought a number of issues of operating a mainframe. As a matter of fact, only now the mainframe has become distributed and more difficult to manage than ever before. On the other hand, all in all, the advantages seem to be more important than the additional complication (Trireme, 2004; Nemeth and Sunderam, 2002). Distributed computing environment is, theoretically, more flexible, open, speedy, and scalable, and it is essential to run and manage a great deal more global and distributed business operations. Additionally, distributed computing environment is believed to be a set of computers that do not share their memory as they share multiprocessors. In other words, the terminals must share resources or data by passing messages. It shows that the object-oriented model can be used for modeling such systems. In this scenario, objects work as a natural symbol for the integration of data and control and, as a result, are considered as the natural units of distribution. In view of the fact that there exists an implicit parallelism, thus, the performance of these systems is enhanced. Furthermore, costly, under-utilized peripherals and resources, for instance, plotters and printers, can be shared to decrease the overall expenses (Trireme, 2004; Nemeth and Sunderam, 2002). Moreover, if a distributed environment is effectively designed, there may be no single reason of a breakdown. In reality, it can be possible to copy services with the intention that maximum accessibility and better flexibility can be ensured. Additionally, more systems can be linked as well as the system can be reconfigured little by little, to decrease the overall expenses or smooth the progress of upgrades to a certain extent like component hi-fi systems. Thus, these and a number of other advantages must be kept in mind against the extra complexity of maintaining a complicated architecture as well as the greater complexity of identifying and understanding such complexity. Moreover, to maintain the infrastructure, compact hi-fi systems are a great deal more sophisticated and easy to install and use as compared to component hi-fi, despite the fact that their quality, power and flexibility are greatly lower (Trireme, 2004; Nemeth and Sunderam, 2002). Utilities of Distributed Computing Support SoftWare Distributed computing environment can be implemented by means of traditional computing environments, for instance, parallel virtual machine (PVM) and some implementations of message passing interfaces (MPIs) or with up-and-coming software frameworks known as computational grids. No matter, whatever the scenario is, the basic goal is to provide support for communication, manage processes and resources, or provide other ways of communication. On the whole, provide a logical view to the application by virtually integrating geographically dispersed resources into an abstract machine. In the above discussed scenarios, the architecture of an application is very similar. However, the intended processes have need of some resources that must be considered at some computational nodes. Additionally, if all the intended process gets all the required resources, it can perform the computation. Consequently, both the traditional and grid systems are similar and any differences are technical ones. One of the major dissimilarities is in how they obtain the resources or how they set up a virtual, imaginary machine from the accessible resources (Nemeth and Sunderam, 2002). This section outlines some of the important software systems that can be used for distributed computing systems. There are numerous applications and physical resources that are shared in a gird computing arrangement. Additionally, the resources that can be shared in distributed computing systems (in grid and clouds) comprise (ITU, 2009): Physical resources The computational power resource sharing Storage devices sharing in distributed arrangement Communication capacity sharing in distributed computing arrangement Virtual resources In a distributed computing environment there are certain virtual resources that work in cooperation to run and manage the distributed computing arrangements. In this scenario, distributed virtual resources can be exchanged as well as are independent of physical location; similar to: Operating systems for running distributed computing arrangement Virtual memory for storage management of distributed computing arrangement Tasks and applications for executing and managing operations of the distributed computing arrangement Software and licenses for distributed computing arrangement Services for handling and managing distributed computing arrangement Layered Structure of Distributed Computing Software This section discusses the layered structure of distributed computing environment. In this scenario, Figure 1 shows a potential structure of the distributed computing arrangement. Such kind of systems is established for other types of distributed computing arrangement, for instance, particularly for grids and clouds, and in a basic stack with 3 layered structures (ITU, 2009): Application layer Mediator (or resource interconnector) Connectivity layer (or shared resources) Figure 1. Stack of a Distributed System. Image Source: (ITU, 2009) In this structure we also have a client layer that is used to show data and information received by the client as input, and response of the other upper system layers. This layer can be developed through a Matlab computing client, web browser or using an Oracle database client that offers some of the other applications that can be addressed (ITU, 2009). Similarly, this software structure of the distributed computing arrangement involves the network and transparent layers that do not depend on the middleware layer. This layered structure performs a mediating function. Therefore, it establishes connection between clients and provisioned and requested technology based resources, balances high working loads among a variety of technology resources and clients. Additionally, this middle layer controls access to limited resources (for example, handling out time on a supercomputer), handling all kinds of tasks, collecting distributed computing statistics, which can afterward be utilized for system and billing management. However, the software based middleware has to be dependable as well as for all time accessible. In fact, it provides communication interface to upper and lower network layers that can be used by the programmers to sketch the system from their requirements. What is more, these distributed computing application interfaces facilitate the system to be extensible, scalable and to manage peak loads, for example, all through the holiday season (ITU, 2009; Topcuoglu et al., 1997). Software Architecture The fundamental idea behind the design of a distributed computing system is to offer a broad software development atmosphere that can be used to construct and run wide ranging applications of a technology based network of diverse resources. Accordingly, distributed computing arrangement is made up of geographically distributed computation stations as demonstrated in Figure 2. In this arrangement, each resource is a one or more distributed computing server. Additionally, at every section of such distributed computing arrangement server executes the server software known as the site manager that is responsible for managing inter-site collaboration and connecting distributed computing environment modules to the online repository (ITU, 2009; Topcuoglu et al., 1997; Berman et al., 1996). Figure 2. Virtual Distributed Computing Environment. Image Source: (Topcuoglu et al., 1997) In addition, the software development framework for network applications can be divided into 3 stages of software development: Specification Stage Application scheduling and configuration Stage Execution and runtime phase The fundamental functionality of these 3 stages is managed by the Distributed Computing Application Editor, Software Scheduler and distributed computing structure demonstrated in the Figure 3 (Topcuoglu et al., 1997; Berman et al., 1996): Figure 3. Interactions among the VDCE Modules. Image Source: (Topcuoglu et al., 1997) Distributed Computing Application Characteristics Various researches show that not every software application is appropriate for a distributed computing structure. In fact, a software application closer to run in real time arrangement will be less appropriate for the distributed computing arrangement. Even though processing jobs that usually take an hour or two cannot derive a great deal advantage if the collaboration and communications among distributed computing systems and the continually changing accessibility of processing users turn out to be a challenge. Additionally, we should consider regarding jobs that take days, hours, months or weeks. Thus, mainly suitable software, as stated by Entropia, composed of non sequential and loosely coupled jobs those are in a batch process by means of an extensive compute-to-data percentage. The extensive computation regarding data ratio moves hand in hand with a high level computation to communications percentage; as there is no need to bog-down the network by transmitting huge volume of data to every user, in various situations we are able to perform so all through the off hours arrangement. Consequently, the software applications with heavy databases that be easily parsed for distribution are suitable for distributed computing environment (Erlanger, 2002; Dasgupta et al., 1991; Hussein et al., 2004). Without a doubt, some applications with individual jobs that require access to huge data sets will be more suitable for better systems as compared to individual personal computers. Besides, when large terabytes of data are concerned, a supercomputer builds a sense as communications can happen across the system’s extremely high speed backplane exclusive of reducing down the communication network. In addition, distributed dedicated system and server clusters will be more suitable for other extremely less data exhaustive applications. Moreover, for a distributed computing application with many personal computers, the necessary data should fit extremely contentedly in the personal computer memory, with much room to spare (Erlanger, 2002; Hussein et al., 2004). Distributed Computing Software Types This section outlines some of the important types of distributed computing arrangement that are beyond the extremely well-known poster SETI@Home software, below given situations are some instances of other kinds of software and application jobs that can be kept in mind to get benefit from a distributed computing arrangement (Erlanger, 2002; Hussein et al., 2004): Distributed computing systems make use of query search software working on a massive database that can be divided all through the plenty of desktop systems. In this way, we can be able to submit query running in a parallel way against every piece of desktop system. Distributed computing systems also require some simulation and complex modeling methods that are used to augment the precision of result by improving the quantity of random trials would as well be suitable. In this scenario, trials could be executed in a parallel way on a number of desktop systems, and combined to attain superior statistical impact. In fact, it is a very useful technique that is utilized in a variety of financial risk analysis areas. Distributed computing systems need some exhaustive search techniques that necessitate searching all through a large number of results or outputs to discover some solutions to problems as well make sense. Consequently, drug screening is one of the most important examples of such kind of software. Distributed computing systems also need a great deal additional computing power. Hence, there are also some application needs of high-tech process management for a large number of users using the system. Distributed computing systems can also involve some weather forecasting, complex financial modeling and geophysical examination software types those are presently very important for handling and managing other systems. Distributed Computing Operating Systems Distributed computing operating systems have several aspects in centralization of tasks and resources. However, these applications vary in a number of ways. For instance, project like Folding@home1, is one of the major examples of distributed computing operating system which is completely a different kind of operating system. Additionally, Linux is also used for the distributed nature of resource management, sharing and management. Furthermore, modified Linux operating system kernel is implementation of management of the high-level distributed computing services. Therefore, distributed computing services that are incorporated in new modified version of Linux are (Hussein et al., 2004; Dasgupta et al., 1991): Process check pointing Facilitate fair scheduling of resources in the Linux kernel VFS interface for customized file systems Distributed Computing Services Management Software For demonstrating the working of distributed computing services we have an application that is known as CloudStack. This application forms the basis for infrastructure clouds and distributed network data center. In this way, the manager of a distributed data center can communicate with operators more quickly and easily for distributed services in their existing infrastructure to present a flexible and on demand distributed services (CloudStack, 2012). Figure 4. Distributed Computing Services Management Software. Image Source: (CloudStack, 2012) Distributed Computing Computational, Storage, and Networking Resources management Software This section discusses another distributed software application computing computational, storage, and networking resources management. In this scenario, OpenStack offers a wide variety of distributed services for cloud operating system. OpenStack offers controls for large scale pools of jobs for managing the storage, computational and networking resources all through a data center. Such services are managed through a dashboard that provides administrators control and allow their clients provision of technology based resources by means of a web based interface (OpenStack, 2012). Figure 5. Compute, storage, and networking resources management Software. Image Source: (OpenStack, 2012) Distributed Computing Application Development Software This section outlines the Distributed Computing Application Development Software ALiCE. Basically, ALiCE is a grid computing project. Alice (Adaptive and scaLable Internet-based Computing Engine) is a lightweight distributed computing application for grid computing middleware for combined computational technology based resources management. In addition, ALiCE offers AOPT and object based programming template library support and services for the fast development of grid software (National University of Singapore, 2003). Distributed Computing Security Management Software Condor-G is a distributed computing security management system that improves the current developments in 2 distinct areas (Wise, 2012): Security and resource access management Administration of computation as well as harnessing of resources in a single management domain Condor-G integrates the inter domain technology based resource administration network protocols of the Globus Toolkit, the intra-domain resource and task management techniques of Condor to permit the client to put together multi domain technology based resources as if they all connected to one personal working domain (Wise, 2012). Distributed Software Data Movement Software Globus is distributed data movement software that can be used in a distributed computing arrangement for the sake of moving large size data securely as well as reliably between distributed and local computing resources. The Globus Toolkit provides a basic distributed computing software application for developing distributed grids that permits distributed computing storage resources, power, technology and scientific instruments to be shared in a protected way inside an institution, corporate or geographic boundaries (Globus, 2012). Distributed Computing Resources Management Software GridSim is a next generation software application for the distributed computing environment that helps resolve large scale problems of engineering, science and commerce. It is normal to entail millions of resources scattering all through a wide range of business, strategies and administrative domains. In addition, the administration and preparation of technology resources in large scale distributed systems is composite and, as a result, it demands refined tools for analyzing as well as fine-tuning the algorithms previous to implementing them to the real systems (Buyya, 2008). Distributed Computing Remote Access Software ZetSolve, or GridSolve, is also a distributed computing software application that has been developed to bring together different technology related computational resources linked with computer networks. Basically, it is an RPC supported distributed agent/client/server software that permits people to remotely access and connect with both software and hardware components (University of Tennessee, 2012). Distributed Computing Multimedia Access Management Toolkit The Access Grid Toolkit is developed by the United States Department of Energy’s Argonne National Laboratory that provides an excellent support for the development of software applications to share audio, video, data and text for real-time association among people at different locations all through the world (means in a distributed computing arrangement) (Mullins, 2006). Conclusion At the present, distributed computing environment is seen everywhere. Indeed, the majority of firms are making effective use of this useful technology to gain business advantages. The distributed environment is in some sense overcoming the limitations of traditional computing offering independence from PCs to the golden or dark period of the mainframe. It is an admitted fact that distributed computing environment is, theoretically, more flexible, open, speedy, and scalable and it is essential to run and manage a great deal more global and distributed business operations. This paper has presented a detailed analysis of some of the important aspects that are related to the distributed computing environment. This paper has discussed a wide variety of software applications that can be used to support a true distributed computer system. This paper has discussed the application structure, working and operational framework in the distributed computing arrangement. References Ahmad, I. and Lau, F., 1999. Special issue on software support for distributed computing. Journal of Parallel and Distributed Computing - Special issue on software support for distributed computing, 59(2), pp.101-06. Berman, F.D. et al., 1996. Application-level scheduling on distributed heterogeneous networks. In Supercomputing '96 Proceedings of the 1996 ACM/IEEE conference on Supercomputing (CDROM)., 1996. IEEE Computer Society Washington, DC, USA. Buyya, R., 2008. GridSim: A Grid Simulation Toolkit for Resource Modelling and Application Scheduling for Parallel and Distributed Computing. [Online] Available at: http://www.cloudbus.org/~raj/gridsim/ [Accessed 29 July 2012]. CloudStack, 2012. CloudStack - Open Source Cloud Computing Software. [Online] Available at: http://www.cloudstack.org/software.html [Accessed 29 July 2012]. Dasgupta, P., LeBlanc, R.J., Ahamad, M. and Ramachandran, U., 1991. The Clouds Distributed Operating System. Computer, 24(11), pp.34-44. Erlanger, L., 2002. Distributed Computing: An Introduction. [Online] Available at: http://www.pcmag.com/article2/0,2817,11769,00.asp [Accessed 30 July 2012]. Globus, 2012. Globus Online: Use The Grid. [Online] Available at: http://www.globus.org/ [Accessed 29 July 2012]. Hussein, N., Kolivas, C. and Haron, F., 2004. Extending The Linux Operating System For Grid Computing. [Online] Available at: http://www.cn.apan.net/cairns/NRW/23-Nur%20Hussein.pdf [Accessed 01 August 2012]. ITU, 2009. Distributed Computing: Utilities, Grids and Clouds. [Online] Available at: http://www.itu.int/dms_pub/itu-t/oth/23/01/T23010000090001PDFE.pdf [Accessed 30 July 2012]. Mullins, R., 2006. U.S. government offers grid computing tool kit. [Online] Available at: http://www.computerworld.com/s/article/9002659/U.S._government_offers_grid_computing_tool_kit [Accessed 03 August 2012]. National University of Singapore, 2003. ALiCE Grid Computing Project. [Online] Available at: http://www.comp.nus.edu.sg/~teoym/alice.htm [Accessed 02 August 2012]. Nemeth, Z. and Sunderam, V., 2002. A Comparison of Conventional Distributed Computing Environments and Computational Grids. Springer-Verlag Berlin Heidelberg. OpenStack, 2012. OpenStack: The Open Source Cloud Operating System. [Online] Available at: http://www.openstack.org/software/ [Accessed 30 July 2012]. Topcuoglu, H. et al., 1997. The Software Architecture of a Virtual Distributed Computing Environment. In HPDC '97 Proceedings of the 6th IEEE International Symposium on High Performance Distributed Computing., 1997. IEEE Computer Society Washington, DC, USA. Trireme, 2004. Distributed computing, middleware and migration. [Online] Available at: http://www.trireme.com/whitepapers/books/oo_methods/oom3e04.pdf [Accessed 01 August 2012]. University of Tennessee, 2012. NetSolve/GridSolve. [Online] Available at: http://icl.cs.utk.edu/netsolve/ [Accessed 30 July 2012]. Wise, 2012. Condor High Throughput Computing. [Online] Available at: http://research.cs.wisc.edu/condor/condorg/ [Accessed 03 August 2012]. Read More