Search

In cloud computing, heterogeneous resources located in different data centers are used to provide pay-per-use performance. Resource provisioning encounters problems due to uncertainty and dispersion of resources in cloud computing environments. According to our research, there are various techniques, but the existing methods and frameworks cannot handle the behaviors of these applications, environment, and resources. The cloud provider wants a framework to execute the task for achieving cost-effective delivery and reliable cloud services. Cloud workload's essential requirement is Quality of Service (QoS) based on adequate resource provisioning as per cloud workload needs. The resource discovery and best paring of workload-resource as per the cloud user's QoS need for workload requirements is an optimization problem. The resource provisioning is appropriate and can be acceptable if offered as per QoS needs in a controlled way. A mutual contract of Service Level Agreement (SLA) is signed between Cloud Service Provider (CSP) and cloud users, which determines the Service Level Objectives (SLO). The SLA contains the QoS parameters and notifies the dependent SLA's recent status. The user will get Quality of Service (QoS), based on prediction, where the resource would manage efficiently and lead to perform loads and deliver cost-effective and reliable cloud services. This paper presents a prediction-based resource management technique called Predictive Cloud Computing Systems (PCCSs). Focus on the self-healing-based prediction that handles unexpected failures and self-configuration-based prediction of resources and applications. Predictive Cloud Computing Systems (PCCSs) performance evaluated in the cloud simulator. The simulation results reveal that Predictive Cloud Computing Systems (PCCSs) achieve better results than existing techniques as execution time, cost-effective, resource conflict, SLA breach while delivering reliable services. [JJCIT 2021; 7(2.000): 206-222]

Summary: Stream processing is a parallel paradigm used in many application domains. With the advance of graphics processing units (GPUs), their usage in stream processing applications has increased as well. The efficient utilization of GPU accelerators in streaming scenarios requires to batch input elements in microbatches, whose computation is offloaded on the GPU leveraging data parallelism within the same batch of data. Since data elements are continuously received based on the input speed, the bigger the microbatch size the higher the latency to completely buffer it and to start the processing on the device. Unfortunately, stream processing applications often have strict latency requirements that need to find the best size of the microbatches and to adapt it dynamically based on the workload conditions as well as according to the characteristics of the underlying device and network. In this work, we aim at implementing latency‐aware adaptive microbatching techniques and algorithms for streaming compression applications targeting GPUs. The evaluation is conducted using the Lempel‐Ziv‐Storer‐Szymanski compression application considering different input workloads. As a general result of our work, we noticed that algorithms with elastic adaptation factors respond better for stable workloads, while algorithms with narrower targets respond better for highly unbalanced workloads. [ABSTRACT FROM AUTHOR]

High Performance Computing (HPC) clouds need to be efficiently shared between selfish tenants having applications with different resource requirements and Service Level Objectives (SLOs). The main difficulty relies on providing concurrent resource access to such tenants while maximizing the resource utilization. To overcome this challenge, we propose Merkat, a market-based SLO-driven cloud platform. Merkat relies on a market-based model specifically designed for on-demand fine-grain resource allocation to maximize resource utilization and it uses a combination of currency distribution and dynamic resource pricing to ensure proper resource distribution among tenants. To meet the tenant’s SLO, Merkat uses autonomous controllers, which apply adaptation policies that: (i) dynamically tune the application’s provisioned CPU and memory per virtual machine in contention periods, or (ii) dynamically change the number of virtual machines. Our evaluation with simulation and on the Grid’5000 testbed shows that Merkat provides flexible support for different application types and SLOs and good tenant satisfaction compared to existing centralized systems, while the infrastructure resource utilization is improved. [ABSTRACT FROM AUTHOR]