Static Scheduling of Moldable Streaming Tasks With Task Fusion for Parallel Systems With DVFS

NASLMRP: Design of a Negotiation Aware Service Level Agreement Model for Resource Provisioning in Cloud Environments

“…This evaluation resulted in the calculation of DHR in relation to the count of tasks executed in total is provided in Table 2 displayed below. The evaluation results depicted in Figure 4, demonstrate that the proposed model exhibits a 4.8% higher DHR than DNNS [9], 4.5% higher DHR than NPS CS [13] , and 5.9% higher DHR than EES [14], which makes it useful for deadline aware scheduling considerations. This results from being included of resource capacity with task deadlines and incremental performance tuning, which assists in identification of optimal schedulingconfigurations.…”

“…Verifying the model's performance, it was assessed based on its task make-span delay (D), computational cycles needed (CC), deadline hit ratio (DHR), and energy efficiency (E) under various VM and task settings. This performance was evaluated in comparison to DNNS 9 , NPS CS 13 , and EES 14 The proposed model, NASLMRP was validated using the Cloud Sim simulator, which aided in the formation of virtual machines (VMs or resources), job scheduling procedures, and assessment of the performance of various model parameters. These datasets were pooled to provide 1 million jobs, which were tested on 400 virtual machines with standard setups.…”

“…These datasets were pooled to provide 1 million jobs, which were tested on 400 virtual machines with standard setups. Based on this assessment technique, the average make-span delay (D) for various Number of Tasks (NT) can be observed from Table 1 After evaluating the proposed model and referring to figure 3, it can be concluded that the model exhibits a performance improvement of 38.5% when compared to DNNS [9] , 46.1% faster performance than NPS CS [13] , and 45.5% faster performance than EES [14], which makes it useful for lowdelay scheduling scenarios. As a result of incorporation of resource capacity and task make-span levels along with incremental performance tuning, which assists in identification of optimal scheduling configurations…”

“…Based on this evaluation, computational cycles (CC) needed for execution is tabulated in table 3. as follows, Based on this assessment and figure 6, it can be seen that the proposed model uses 24.5%, 18.3%, and 16.5% less computational energy than DNNS [9], NPS CS [13], and EES [14], respectively, making it suitable for low energy scheduling scenarios. This is because incremental performance tuning combined with the incorporation of resource capacity, RAM, MIPS, and the number of VMs aids in the identification of the best scheduling configurations.…”

Workflow Simulation and Multi-Threading Aware Task Scheduling for Heterogeneous Computing

“…An extensive comparison of several TS algorithms for moldable tasks is carried out in [17], for identical processors (both theoretical and experimental). [28] employs ILP for streaming applications while [39] employs ILP considering inter intra task communications. Last, in [2] they present efficient algorithms for scheduling an application on hybrid platforms of identical CPUs and GPUs.…”