The model simulation is aimed at finding the emulator which offers the best output in the shortest time with maximum utility gained from processing the data or information. The evaluation showcases that for a centrally organized queuing simulator uses a single global queue of ready which is usually available to all processors in the system. Nevertheless, it is not appropriate for use since it allows a mutually exclusive access which throttles when there is an increase in number of ruing process in the simulation. This I turn slows down the system; hence it is not advisable for use on large parallel systems.
/>
The beneficial use of distributes or hierarchical organized systems offer better queuing for processes in a simulation environment. The distributed organizations simulations are based on the local ready queues which are linked with the processors. This evades the problematic issue of ready queue access contention. However, this presents other hitches. The foremost problematic issue with the distributed organization being to find a suitable ready process queue for every new arrival task.
This problematic issue brings forth the need for a more efficient processing queue technique through simulation to support the fast execution of instructions and carrying out of processes in the machines. The objective presents a greater system, that is, the hierarchical organization, which delivers a model alike to the centralized organization even when there is no access contention similar in the distributed organization.
The performance analysis highlights and proves that the hierarchical organization offers a better solution. This is concluded from a simulation that presents the output parameters form the impact of the system size and the access contention. This further details the response time and utilization of processes. The inference demonstrates that the simulation processes with the shortest time and maximum utility depict the best queuing system.
The changes that I would recommend for the study that the authors carried out are the alteration in the memory used in the simulation examination or experimentation. Since the simulation focused on shared memory, recommending the use of dedicated memory for the systems would be great at singling out processes and queuing them to specified memory for constant use. Dedicated memory speeds up the system compares to the use of shared memory; this should be a section of the study the researcher s would have looked into for improvement.
The queuing simulations that involve the use of hierarchical organized systems include their use in an ATM. Here the ATM waiting lines are lined up for task execution in the categorized formation. The other instance of the use of the simulations is in the discrete event systems (DES). These comprise of dynamic systems which advance in time by the incidence of actions at conceivably uneven time intermissions. Instances of DES comprise of computer-communications systems, flexible manufacturing systems, traffic systems, production lines, flow networks, and coherent lifetime systems.
The readings on event simulation offer a snapshot of how the real world events are supposed to occur. This gives us the knowledge of how events happen, especially the unseen (intangible) ones, thus gives us the conclusive knowledge approach to certain occurrences in a system. For example, when a computer node operation slows down in a network like on an ATM machine, an experienced individual can conclude that maybe there is an overload in the system processes. The hands-on experience also derives the feel of satisfaction for the realistic practicality in the operating environment.
References
Kostin, A., & Ilushechkina, L. (2010). Modeling and simulation of distributed systems. Singapore: World Scientific.
Serrazi, G. (2007) Shared-Memory Multiprocessor Systems—Hierarchical Task Queue. University of Lugano, Advanced Learning and Research Institute. Retrieved December 4, 2011, fromhttp://www.idsi.md/files/file/publicatii/PERFORMANCE EVALUATION_Exercises_Zamsa.pdf
Menascé, D. A., Almeida, V. A. F., & Dowdy, L. (1994). Capacity planning and performance modeling: From mainframes to client-server systems. Englewood Cliffs, N.J: PTR Prentice Hall.
