In Cloud systems, Virtual Machines (VMs) are scheduled to hosts according to their instant resource usage (e.g. to hosts with most available RAM) without considering their overall and long-term ...utilization. Also, in many cases, the scheduling and placement processes are computational expensive and affect performance of deployed VMs. In this work, we present a Cloud VM scheduling algorithm that takes into account already running VM resource usage over time by analyzing past VM utilization levels in order to schedule VMs by optimizing performance. We observe that Cloud management processes, like VM placement, affect already deployed systems (for example this could involve throughput drop in a database cluster), so we aim to minimize such performance degradation. Moreover, overloaded VMs tend to steal resources (e.g. CPU) from neighbouring VMs, so our work maximizes VMs real CPU utilization. Based on these, we provide an experimental analysis to compare our solution with traditional schedulers used in OpenStack by exploring the behaviour of different NoSQL (MongoDB, Apache Cassandra and Elasticsearch). The results show that our solution refines traditional instant-based physical machine selection as it learns the system behaviour as well as it adapts over time. The analysis is prosperous as for the selected setting we approximately minimize performance degradation by 19% and we maximize CPU real time by 2% when using real world workloads.
Optimizing the virtual machine (VM) migration is an important issue of server consolidation in the cloud data center. By leveraging the content similarity among the memory of VMs, the time and the ...amount of transferred data in VM migration, as well as the pressure of network traffic, can be reduced. There are two problems in server consolidation: (1) determining which VMs should be migrated from the overloaded hosts (VM selection problem) and (2) how to place these VMs to the destination hosts (VM placement problem). By exploiting the content similarity, we redefine the above two problems into one problem to minimize the transferred memory data in VM migration. Given a fixed host overloaded threshold, an approximation algorithm is proposed to solve the problem with one overloaded host and one destination host. For the case of multiple overloaded hosts and destination hosts, two heuristic algorithms are presented with fixed and dynamic overloaded threshold respectively. We conduct a real workload trace based simulation to evaluate the performance of our algorithms. The result shows that our algorithms can produce fewer transferred VM memory data and consume less energy than existing policies.
•CVSP is defined to minimize the amount of transferred VM memory pages.•An approximation algorithm for a special case of CVSP.•Two heuristic algorithms with fixed and dynamic overloaded thresholds.•A real workload trace-driven simulation is conducted to evaluate the performance.
To improve resource utilization and energy efficiency, cloud data centers use virtual machine (VM) consolidation to consolidate VMs to fewer physical machines (PMs) through live VM migration. ...However, improper VM placement may cause frequent VM migrations and constant on–off switching of PMs, which results in lower service quality and increased energy consumption. In this paper, we address this problem by proposing an effective and efficient VM consolidation approach called EQ-VMC, which has the goal of optimizing energy efficiency and service quality. In our approach, a discrete differential evolution algorithm is developed to search for the global optimum solution for VM placement. By integrating this solution with a set of algorithms proposed for effective host overload detection, VM selection, and under-loaded host detection, EQ-VMC effectively reduces energy consumption and improves quality of service (QoS). Extensive simulation demonstrates its effectiveness and shows its superiority to previous VM consolidation methods.
•First, we note that VM placement is an authentic combination optimization issue with multiple resource constraint.•Then, the probable mappings between VMs and PMs are abstracted as a piece of limited search space, and which corresponds to a population of heuristic evolutionary algorithm. Each individual of population is identical to a real mapping between VMs and PMs during a cycle of VMs consolidation.•Next, we define a combination optimization model for handling VM placement to achieve the optimal mapping between VMs and PMs in the search space. The solution of the optimization model is performed by an improved heuristic evolutionary algorithm to guarantee the globally optimal results, namely, the optimum VM placement scheme.•At last, the proposed EQ-VMC method integrates sub-algorithms on host overloading detection, VM selection and under-loaded host detection for VMs consolidation. Comparison and validation are performed using the CloudSim toolkit. The experimental results show that the presented EQ-VMC method is promising in degrading energy consumption and host overloading risk, as well as in improving QoS. Thereby its effectiveness and efficiency have been validated.
•We determine the upper limit and lower limit of physical machine utilization.•We propose trusted virtual machine migration rule.•The HTVM2 algorithm is proposed and validated based on the above ...rule.
With the continuous development and maturity of cloud computing technology, the scale and number of cloud data center (CDC) are also expanding. This increasingly draws attention to the problem of high energy consumption in CDCs. Dynamic virtual machine (VM) consolidation is a promising approach for reducing energy consumption. VM migration, as a VM consolidation technology, can effectively improve the utilization of physical machine (PM) and optimize the scheduling process of CDCs. However, most VM integration algorithms, in existing research, are aimed at improving the utilization of PMs. Excessive utilization of PMs may increase the competition for shared resources among the VMs running on them. As a result, the performance of these VMs deteriorates, and the execution time of cloud tasks is increased or even interrupted. This study systematically analyzes the overall architecture of CDCs. Subsequently, migration rules are established for the one-dimensional and multidimensional trusted VMs. A high- applicability heterogeneous CDC resource management algorithm based on trusted VM migration (HTVM2) is then proposed. The proposed algorithm not only solves the one-dimensional VM migration problem of homogeneous and heterogeneous CDCs but also those of multi-dimensional VMs. This improves the success rate of VM migration, reduces the energy consumption of the CDC, and improves load balancing while ensuring VM performance. Finally, the algorithm was compared with the other three algorithms outperforming them all, as demonstrated by experimental results.
Infrastructure-as-a-service (IaaS) Clouds concurrently accommodate diverse sets of user requests, requiring an efficient strategy for storing and retrieving virtual machine images (VMIs) at a large ...scale. The VMI storage management requires dealing with multiple VMIs, typically in the magnitude of gigabytes, which entails VMI sprawl issues hindering the elastic resource management and provisioning. Unfortunately, existing techniques to facilitate VMI management overlook VMI semantics (i.e at the level of base image and software packages), with either restricted possibility to identify and extract reusable functionalities or with higher VMI publishing and retrieval overheads. In this paper, we propose Expelliarmus, a novel VMI management system that helps to minimize VMI storage, publishing and retrieval overheads. To achieve this goal, Expelliarmus incorporates three complementary features. First, it models VMIs as semantic graphs to facilitate their similarity computation. Second, it provides a semantically-aware VMI decomposition and base image selection to extract and store non-redundant base image and software packages. Third, it assembles VMIs based on the required software packages upon user request. We evaluate Expelliarmus through a representative set of synthetic Cloud VMIs on a real test-bed. Experimental results show that our semantic-centric approach is able to optimize the repository size by 2.3−22 times compared to state-of-the-art systems (e.g. IBM’s Mirage and Hemera) with significant VMI publishing and slight retrieval performance improvement.
•A novel semantic model representing VMIs as structured graphs.•VMI clustering based on functionality with low similarity computation overheads.•A semantics-aware VMI decomposition method without costly content deduplication.•Optimized VMI assembly with compatible base image and selective package retrieval.•Reduced repository size upto 22 times with improved VMI publishing and retrieval.•Scalability analysis for increase in repository size and VMI retrieval performance.
Virtual machine (VM) packing plays an important role in improving resource utilization in cloud data centers. Recently, memory content similarity among VM instances has been used to speed up multiple ...VM migration in large clouds. Based on this, many VM packing algorithms have been proposed, which only considered the memory capacity of physical machines (PMs) as the resource constraint. However, in practice the results of such algorithms are not feasible, because thy may not satisfy the constraints of multiple resources (e.g., CPU of the PMs). Besides, the granularities of memory sharing in existing studies are very coarse, and they cannot fully leverage the benefits of memory content similarity which mainly appears at memory page level. In this paper, we study the page-sharing-based VM packing that considers constraints in multiple resources. Given a set of VM instances that share a large number of common memory pages, we pack them into the minimum number of PMs, subject to the constraints in the multiple resources on the PMs. This problem is solved in two steps. First, we pack the maximum number of VMs into a given PM, and then propose an approximation algorithm. The approximation ratio is better than that of the existing algorithm. Then, based on this approximation algorithm, we propose a heuristic algorithm to solve the general problem. Experimental results show that our heuristic algorithm outperforms existing approaches with at most 25% less required PMs and at most 40% less memory page transferring.
•We study the page-sharing-based VM packing with multiple constraints.•An algorithm with better approximation ratio is proposed under special case.•For the general case of the problem, a heuristic algorithm is proposed.•Simulation results show the efficiency of our heuristic algorithm.
The large-scale virtualized Cloud data centers consume huge amount of electrical energy leading to high operational costs and emission of greenhouse gases. Virtual machine (VM) consolidation has been ...found to be a promising approach to improve resource utilization and reduce energy consumption of the data center. However, aggressive consolidation of VMs tends to increase the number of VM migrations and leads to over-utilization of hosts. This in turn affects the quality of service (QoS) of the applications running in the VMs. Thus, reduction in energy consumption and at the same time ensuring proper QoS to the Cloud users are one of the major challenges among the researchers. In this paper, we have proposed an energy efficient and QoS-aware VM consolidation technique in order to address this problem. We have used Markov chain-based prediction approach to identify the over-utilized and under-utilized hosts in the data center. We have also proposed an efficient VM selection and placement policy based on linear weighted sum approach to migrate the VMs from over-utilized and under-utilized hosts considering both energy and QoS. Extensive simulations using real-world traces and comparison with state-of-art strategies show that our VM consolidation approach substantially reduces energy consumption within a data center while delivering suitable QoS.
Virtual machine consolidation aims at reducing the number of active physical servers in a data center so as to decrease the total power consumption. In this context, most of the existing solutions ...rely on aggressive virtual machine migration, thus resulting in unnecessary overhead and energy wastage. Besides, virtual machine consolidation should take into account multiple resource types at the same time, since CPU is not the only critical resource in cloud data centers. In fact, also memory and network bandwidth can become a bottleneck, possibly causing violations in the service level agreement. This article presents a virtual machine consolidation algorithm with multiple usage prediction (VMCUP-M) to improve the energy efficiency of cloud data centers. In this context, multiple usage refers to both resource types and the horizon employed to predict future utilization. Our algorithm is executed during the virtual machine consolidation process to estimate the long-term utilization of multiple resource types based on the local history of the considered servers. The joint use of current and predicted resource utilization allows for a reliable characterization of overloaded and underloaded servers, thereby reducing both the load and the power consumption after consolidation. We evaluate our solution through simulations on both synthetic and real-world workloads. The obtained results show that consolidation with multiple usage prediction reduces the number of migrations and the power consumption of the servers while complying with the service level agreement.