Skip to main content

Advances, Systems and Applications

Journal of Cloud Computing Cover Image

Table 1 Comparison between the multi-factor related work

From: Multi-Dimensional Regression Host Utilization algorithm (MDRHU) for Host Overload Detection in Cloud Computing

Related work Parameters Target Algorithm name Advantages Disadvantages
Tian et.al [16] CPU, memory and network BW Resource scheduling Dynamic and Integrated Resource Scheduling (DAIRS) -It provides a dynamic and integrated resource scheduling which treats CPU, memory and network BW integrated for both physical machines and virtual machines -It provides integrated measurement of the total imbalance level of a Cloud data center as well as the average imbalance level of each server for performance evaluation It did not consider energy-efficiency for Cloud data centers
Tang et.al [17] CPU, Communication network in a data center Virtual Machine Placement in server consolidation Hybrid GA (HGA) -It considers the energy consumption in both physical machines and the communication network in a data center -Compared with existing heuristic algorithms, the HGA can generate much better solutions as the HGA is basically a global search algorithm while heuristic algorithms are local search algorithms, which may be trapped at a local optimum during their search process Its computation time is higher than other existing heuristic algorithms
Castro et.al [18] CPU and RAM Virtual machine Placement CPU and RAM Energy aWare (CREW) -It improves VM placement by employing a power model that considers the energy consumed by both CPU and RAM -CREW offers the best trade-off between energy saving and SLA violation It considers CPU and RAM only and it did not consider the communication network in a data center
Li et.al. [19] CPU and disk Host overload/ underload double threshold detection and VM placement Multi-resource double threshold method and Modified Particle Swarm Optimization (MPSO) VM placement - It designed a method of double threshold with multi resource utilization to trigger the migration of VMs. The Modified Particle Swarm Optimization method is introduced into the consolidation of VMs to avoid falling into local optima which is a common defect in traditional heuristic algorithms It did not consider the memory and communication network in a data center
Farahnakian et al. [20] CPU and memory utilizations A virtual machine consolidation A Utilization Prediction-aware VM Consolidation (UP-VMC) approach - UP-VMC goes beyond the existing works which only consider CPU utilization by also considering memory. Combining both memory and CPU utilization, UP-VMC can better identify causes of SLA violations and consequently prevent them from happening. - In contrast to the existing VM consolidation methods which mostly rely on the current resource utilization of PMs, UP-VMC considers both current and future resource utilization It should take into account network resource utilization and traffic to optimize VM placement.
Abdelsamea et al. [11] CPU and RAM and network BW Host overload detection Multiple Regression Host Overload Detection (MRHOD) -It significantly reduces energy consumption while ensuring a high level of adherence to Service Level Agreements (SLA) It uses a workload that is insignificant for BW