Article Preview
TopIntroduction
The real-time system produces a correct result within a specified period and classified into three basic categories, Hard Real-Time, Soft Real-Time, and Firm Real-Time. At any point of instance, the violation of results is not acceptable in a hard real Time System. Typically, a hard real-time system is always having a hard deadline for each task set i.e., and every task must complete before a specified deadline. Automatic Flight Control System, Space Mission Control System, and Nuclear Power Plant are an example of Hard Real-Time system. In the soft and firm real-time system, user compromise with result and time, which means the execution of the task set, is completed after the deadline, and its result may be useful, according to Biondi et al. (2015) and Buttazzo et al. (2012).
Nowadays, for enhancing the performance, the real-time system incorporates multiple processors so that researchers are keen to develop the new type of scheduling algorithm and synchronization protocol. Still, the research is going on in this field relatively single processor scheduling. In the multiprocessor system, the complexity is increasing so that more research needed for simplification. Multiprocessor hard real-time system requires completion of every job within a given period and also achieving logically correct results; otherwise, some of the results would be valueless and of serious failures of the system will occur stated by Baek.et al. (2018) and Brandenburg et al. (2014). Task allocation in the multiprocessor system will be solved under the two categories, partitioned and global scheduling. In partitioned scheduling, the tasks are allocated statically to the processor means a permanent allocation of the task to a processor, and no migration is allowed. In global scheduling, tasks are allocated dynamically to the ideal processor from a single shared task queue. In general global scheduling allows a job to run at any processor by Cheng. et al. (2005) and Hsiu et al. (2011).
K.C. Okafor et al. (2016) proposed a resource allocation and task scheduling approach in the field of distributed cloud computing networks (DCCNs). Inefficient resource allocation is the main reason for network failure in the cloud. The failure of a cloud network that is running a real-time critical service is highly undesirable. The authors observed the problem of scheduling and resource allocation of a user’s workload in the cloud, for that they proposed an optimal cost scheduling algorithm. In this work, the virtual machine (Vm) algorithms were developed to provide good performance in a cloud computing environment. In this efficient algorithm, the tasks are given to the virtual machines for execution based on first come first serve and the deadline of tasks checked for every high and low priority. These works mainly focus on the dynamic allocation of virtual resources based on the intensities of workload in the distributed cloud. They also proposed architecture for DCCNs resource allocation and task scheduling with the virtual machine. Here a cloud coordinator/scheduler initiates a request to the task scheduler after performing computation on data for virtual machines. The CloudSim tool has been used for the implemented network while ensuring that virtual machines allocated as hosts based on the capacity of the cloud service coordinator available. Here the QoS parameters such as network throughput, resource availability, service delay time, and efficient utilization of resources were investigated. The experimental results show that the enhanced server performance reduces the time and offering lesser average waiting time even when the requesting rate goes high. Resources used the most extended duration of time. The result shows that the average responses’ in percentage have abstained in DCCN is 40%, in DCell is 33.33%, and BCube is 26.67%. That indicates that the performance has enhanced by virtualization.