On Manipulating Dynamic Fluctuation Drawbacks In A Virtualized Environment

Recently, different domains intensively direct their business to Cloud Systems for Cloud’s availability, scalability, in addition to the promised cost savings, which is the most attractive advantage since you only pay for what you actually use. However, in Virtualized Environment the workload behavior fluctuates dynamically producing undesirable situations such as load imbalance, lower utilization, and workload hotspot. In this literature, we investigate the current technology and the running researches covering these issues, and then we introduce a new research study to overcome these drawbacks in two levels.In the elemental level, intensive trend to Cloud Systems in addition to nowadays fluky applications (i.e. social networks, web hosting, content delivery) that exploit the power of the Cloud reveal the dynamic changes as noticeable characteristic of Cloud Systems. As a result, the workload management for these resources becomes a complex task. To the best of our knowledge, most researches in this area tried to find different solutions by enhancing different resources scheduling methodologies and some tries to find other solutions by exploiting the power of the current hardware technologies. However in all cases, an obvious solution to keep step with these changes was to frequently update Cloud’s resources to proper the change demands.VMs resources adaptation can be in the form of horizontal scaling (i.e. adding new server replicas and load balancers to distribute load among all available replicas) or vertical scaling (on-the-fly changing of the assigned resources to an already running instance). However, in cases where the application phase behavior is very dynamic these techniques result poor performance and may lead to infrequent peak loads which drive to low average utilization of resources.Fortunately, even the workload fluctuates rapidly, these changes format repeated sequences which can be logged and handled in similar manners as sentences in language modeling. Therefore, we propose a proactive model based on an application behaviors prediction technique to predict the future workload behavior of the virtual machines (VMs) executed at Cloud hosts. The predicted information can help VMs to dynamically and proactively be adapted to satisfy the provider demands in terms of increasing the utilization and decreasing the power consumption; and to enhance the services in terms of improving the performance with respect to the Quality of Services (QoS) requirements and dynamic changes demands.We study the performance of our Proactive model and compare the results with Monitoring Model. For the sake of this, we performed two benchmarks. In the first, the model maintains information about all VMs and the number of requests currently allocated to each VM. In the second benchmark, the model maintains information manipulated by our Proactive model after monitoring. Each benchmark runs for specific time.Firstly, we have tested the proposed model using CloudSim simulator, and the experiments show that our model is able to avoid undesirable situations caused by dynamic changes such as (peak loads, low utilization) and can decrease the losses of energy consumption, overheating, and resources wastage up to45%on average.Then we direct our experiments to more practical environment, by repeating the same approach and create a proactive workload management model for resources in Virtualized Environment to analyze Virtual Machines workload behavior and to adopt adequate scheduling schema and resource assignment in order to enhance the system’s utilization, throughput and response time. We have implemented the proposed model by modifying the existing Xen Cloud Platform and evaluated the performance using representative different benchmarks. Our evaluations show that the proposed model can decrease average waiting time by29%.In wider scope level, Virtual Machine migration is a promising solution to overcome the dynamic fluctuation drawbacks however an algorithm based on single objective, usually the service level agreement (SLA) used to direct the migration. On the contrary, there exist different conflicting goals behind the migration process (i.e. load volume, power consumption, and resource wastage).Indeed migration is a direct solution for the workload issues. However, there are some other important factors to be considered in the migration process such as, utilization of resources to prevent imbalance VMs placement, and migration cost which basically depends on the size of the migrated VMs and the transfer rate; furthermore, power consumption is another important factor because there are other crucial problems that arise from high power consumption as insufficient or malfunctioning cooling system can lead to overheating of the resources reducing system reliability and devices lifetime. In this research, we consider the migration process as a Multi-Objective problem where the objectives are typically non-commensurable. Consequently, we propose a novel migration policy consolidated by a new elastic Multi-Objective Optimization strategy to evaluate different objectives including migration cost simultaneously and to provide the flexibility to facilities different situations. Moreover, we verified the policy by intensive set of experiments using CloudSim simulator and the results ensure the efficiency of the policy to control system performance by adjusting the migration objectives to appropriate different situations.