MClock: handling throughput variability for hypervisor IO scheduling

Virtual Machine (VM) allocation for multiple tenants is an important and challenging problem to provide efficient infrastructure services in cloud data centers. Tenants run applications on their allocated VMs, and the network distance between a tenant's VMs may considerably impact the tenant's Quality of Service (QoS). In this study, we define and formulate the multi-tenant VM allocation problem in cloud data centers, considering the VM requirements of different tenants, and introducing the allocation goal of minimizing the sum of the VMs' network diameters of all tenants. Then, we propose a Layered Progressive resource allocation algorithm for multi-tenant cloud data centers based on the Multiple Knapsack Problem (LP-MKP). The LP-MKP algorithm uses a multi-stage layered progressive method for multi-tenant VM allocation and efficiently handles unprocessed tenants at each stage. This reduces resource fragmentation in cloud data centers, decreases the differences in the QoS among tenants, and improves tenants' overall QoS in cloud data centers. We perform experiments to evaluate the LP-MKP algorithm and demonstrate that it can provide significant gains over other allocation algorithms.

Lightweight container technology has emerged as a fundamental component of cloud-native computing, with the deployment of containers and the balancing of loads on virtual machines representing significant challenges. This paper presents an optimization strategy for container deployment that consists of two stages: coarse-grained and fine-grained load balancing. In the initial stage, a greedy algorithm is employed for coarse-grained deployment, facilitating the distribution of container services across virtual machines in a balanced manner based on resource requests. The subsequent stage utilizes a genetic algorithm for fine-grained resource allocation, ensuring an equitable distribution of resources to each container service on a single virtual machine. This two-stage optimization enhances load balancing and resource utilization throughout the system. Empirical results indicate that this approach is more efficient and adaptable in comparison to the Grey Wolf Optimization (GWO) Algorithm, the Simulated Annealing (SA) Algorithm, and the GWO-SA Algorithm, significantly improving both resource utilization and load balancing performance on virtual machines.

In nowadays cloud computing systems, leveraging the virtualization technology, the customer's requested data computing or storing service is accommodated by a set of mutual-communicated Virtual Machines (VM) in a scalable and elastic manner. These VMs are placed in one or more datacenter nodes according to nodes' capacities, failure probabilities, etc. The VM placement availability refers to the probability that at least one set of the whole customer's requested VMs operates during the entire requested lifetime. The placed VMs should obey the agreed-upon availability, otherwise the cloud provider may face revenue loss. In this paper, we study the problem of placing at most H sets of k requested VMs on minimum number of datacenter nodes, such that the VM placement availability requirement is satisfied and each VM pair has a communication delay no greater than the specified. We prove that this problem is NP-hard. We subsequently propose an exact Integer Nonlinear Program (INLP) and an efficient heuristic to solve this problem. Finally, we conduct simulations to compare the proposed algorithms with two existing heuristics in terms of acceptance ratio, average number of used nodes and running time.

Smart Cloud resource management has gained much research interest in recent years. Optimized resource allocation is necessary while moving from existing computing infrastructure to smart cloud infrastructure. This will allow cloud service providers to get more benefits out of what they already have, increasing capacity and efficiency of existing cloud infrastructure. Optimal Virtual Machine (VM) allocation is a major challenge in cloud resource management. Development of a VM placement strategy which can guarantee required bandwidth and optimizes data transfer delay between VMs is a challenging issue. This paper presents an optimization algorithm to ensure guaranteed bandwidth while minimizing data transfer delay between VMs. Simulation results show that the generated VM placement plan gives minimum inter-distance between allocated VMs while satisfying bandwidth requirements compared to the other existing algorithms.

Virtualization is a rapidly evolving technology that provides a range of benefits to computing systems, such as improved resource utilization and management, application isolation and portability, and system reliability. Among these features, live migration, resources management (including vCPU scheduling and I/O management) are core functions. Live migration of virtual machine (VM) provides a significant benefit for virtual server mobility without disrupting service. It has become an extremely powerful tool for system management in a variety of key scenarios, such as VM load balancing, fault tolerance, power management and other applications. Experimentations and traces show that the performance of live migration is not good enough for different applications. Of course, based on the virtualization architecture management schemes for CPU and I/O resources also need to be reconsidered when supporting different applications with different workload. In this talk, some typical issues in virtualized system will be discussed. Firstly, to take into account the migration overhead in migration decision making, we thoroughly analyze the key parameters that affect the migration cost from theory to practice, and construct two application oblivious models for the cost prediction by using learned knowledge about the workloads at the hypervisor (also called VMM) level. We evaluate the models using five representative workloads on a Xen virtualized environment. Based on the model, we have proposed two live migration schemes for different scenarios: a novel approach that adopts checkpointing/recovery and trace/replay technology to provide fast, transparent VM migration for applications with high reliability, and a memory-compression-based VM migration system for normal applications. Secondly, the asynchronous-synchronous disk I/O model in a typical virtualized system exhibits several problems. For example, when the frontend fails abruptly, the unsaved data in the frontend's cache will be lost. To address the problems, we introduce a new I/O model. In this model, rather than performing the asynchronous-synchronous operations for an asynchronous I/O write request, the frontend file system uses synchronous operations to deal with the I/O request and the backend file system performs asynchronous operations to write the data to the hard disk. A prototype system called HypeGear is implemented on the Xen hypervisor. Thirdly, VMM schedulers have focused on fairly sharing the processor resources among domains, rarely consider VCPUs' behaviors. However, this can result in poor application performance to overcommitted domains if there are concurrent programs hosted in them. We review the properties of both Xen's Credit and SEDF schedulers, and show how these schedulers may seriously impact the performance of the communication-intensive and I/O-intensive concurrent applications in overcommitted domains. A novel approach, that dynamically scales the context switching-frequency by selecting variable time slices according to VCPUs behaviors, is then proposed to improve the Credit scheduler more adaptive for concurrent applications.

Virtualisation enables multiples virtual machines (VMs) to co-locate on a same physical machine with total isolation. Hence using VMs to launch web services or applications is the common trend nowadays in enterprise information technology (IT). Data centre provides infrastructure to create, configure and manage VMs. It has seen as a utility that clients can pay for only as needed. The growing complexity of modern networked computer systems with virtualisation technology necessitates the needs efficient resource management. We have proposed an intelligent resource manager to control the resource allocation in Xen virtualised environment for dynamically allocating resources to individual VM. Our resource management architecture comprises of fuzzy logic based controller. Experimental results shows that with the proposed system data centre can efficiently allocate CPU resources to VMs that have been produced by customers. The scaling of CPU resources is automatically done in accordance with dynamically changing workload at a minimum granularity of 2 seconds. It improves the resource utilisation by 30% as compared to the ideal method while maintaining throughput as equivalent to the ideal workload allocation.

Impact of Factors Affecting Pre-copy Virtual Machine Migration Technique for Cloud Computing

Cloud computing(CC) delivers multiple services to users by processing complex tasks through internet connections to serve as a dynamic, dependable, and flexible computing solution. With the increased Internet speed, the user demands to perform many tasks through computing provided by cloud virtual machines (VMs) created in heterogeneous servers. The complexity of cloud computing(CC) increases through the diversity of servers and the increased demand for resources to perform more tasks, which makes the cloud data center(CDC) unbalanced in load. Therefore, future cloud systems will require more effective resource management(RM) methods to balance load and improve Quality of Service (QoS). This paper proposes KMPPVM-DRM, a dynamic clustering and scheduling approach for VM resource allocation(RA) based on the KMeans++ algorithm. VMs are clustered into three groups (high, medium, low) according to their real-time normalized CPU and RAM weights, using weighting parameters α = 0.6 and β = 0.4. Tasks are categorized into three levels based on length and mapped to appropriate VM clusters. Within each cluster, a Utilization-Level Comparator (ULC) dynamically selects the optimal VM for task execution. The experimental results were conducted using CloudSim, with 5 and 10 VMs and task counts ranging from 1000 to 10000. The proposed model was compared with PSO, ACO, and DBSCAN algorithms using execution time, average start time, and average finish time. Results show that with 10 VMs, the proposed model achieved a harmonic mean execution time of 886.72 ms, compared to 4102.99 ms (PSO), 4672.25 ms (ACO), and 3071.51 ms (DBSCAN). It also attained the lowest harmonic mean start and finish times of 166.41 ms and 167.19 ms, respectively. Relative reduction in execution time against DBSCAN ranged from 69.90% to 71.52%, with start and finish time improvements between 52.65% and 81.51%.To ensure statistical reliability, a paired t-test confirmed that the performance improvements were statistically significant (p < 0.05) across all task sizes. The findings confirm that KMPPVM-DRM enhances resource allocation(RA) efficiency and scheduling effectiveness, maintaining balanced loads even with limited resources (e.g., only 5 VMs) and outperforming PSO, ACO, and DBSCAN in all tested scenarios.

Cloud infrastructures make extensive use of hypervisors (e.g., Xen, ESX), containers (e.g., LXC), and high-level virtual machines (e.g., CLR, Java), broadly known as virtual machine (VM) technologies, to achieve workload isolation and efficient resource management. Isolation is a static mechanism that relies on hardware or operating system support to be enforced. Resource management is dynamic, and VMs must self-adapt or be instructed to adapt in order to fit their guest’s needs. In this chapter, we review the main approaches for adaptation and monitoring in virtual machines deployments, their tradeoffs, and their main mechanisms for resource management. We frame them into an adaptation loop where sensors are monitored (e.g., page utilization), decisions are made (e.g., if-else rule, proportional-integral-derivative controller), and actions are performed using actuators (e.g., share page, change heap size). As is common in systems research, improvement in one property is accomplished at the expense of some other property. So, we present a taxonomy that, when applied to different solutions that use or augment virtual machines, can help visually in determining their similarities and differences. We analyze adaptability in virtual machines using three seemingly orthogonal characteristics: responsiveness (R), comprehensiveness (C), and intricateness (I). The process of classification and comparing systems is detailed, and several representative state of the art systems are evaluated.

A Service Sustainable Live Migration Strategy for Multiple Virtual Machines in Cloud Data Centers

In a shared virtual computing environment, dynamic load changes as well as different quality requirements of applications in their lifetime give rise to dynamic and various capacity demands, which results in lower resource utilization and application quality using the existing static resource allocation. Furthermore, the total required capacities of all the hosted applications in current enterprise data centers, for example, Google, may surpass the capacities of the platform. In this paper, we argue that the existing techniques by turning on or off servers with the help of virtual machine (VM) migration is not enough. Instead, finding an optimized dynamic resource allocation method to solve the problem of on-demand resource provision for VMs is the key to improve the efficiency of data centers. However, the existing dynamic resource allocation methods only focus on either the local optimization within a server or central global optimization, limiting the efficiency of data centers. We propose a two-tiered on-demand resource allocation mechanism consisting of the local and global resource allocation with feedback to provide on-demand capacities to the concurrent applications. We model the on-demand resource allocation using optimization theory. Based on the proposed dynamic resource allocation mechanism and model, we propose a set of on-demand resource allocation algorithms. Our algorithms preferentially ensure performance of critical applications named by the data center manager when resource competition arises according to the time-varying capacity demands and the quality of applications. Using Rainbow, a Xen-based prototype we implemented, we evaluate the VM-based shared platform as well as the two-tiered on-demand resource allocation mechanism and algorithms. The experimental results show that Rainbow without dynamic resource allocation (Rainbow-NDA) provides 26 to 324 percent improvements in the application performance, as well as 26 percent higher average CPU utilization than traditional service computing framework, in which applications use exclusive servers. The two-tiered on-demand resource allocation further improves performance by 9 to 16 percent for those critical applications, 75 percent of the maximum performance improvement, introducing up to 5 percent performance degradations to others, with 1 to 5 percent improvements in the resource utilization in comparison with Rainbow-NDA.

This paper advocates virtualization technology as a methodology to solve the security problems that an organization has to face when contributes with its resources to a grid. In particular, this technology makes it possible to increase the overall security of any system by inserting a set of controls into the code that implements one virtual machine. In this way, a secure cooperation among virtual machine can be implemented. This generalizes the current approach that exploits virtualization only for the confinement of alternative programming environments resulting from the partitioning of a physical machine into a set of non-cooperating virtual machines. The ability to support cooperation among virtual machines may be used to define networks of cooperating virtual machines to execute distributed applications. The paper describes a general purpose approach to security based upon virtual networks of cooperating virtual machines and applies it to one of the most challenging problems: that of securing a grid environment

Though virtual Trusted Platform Module (vTPM) make trusted computing technology compatible with virtualization in individual computers, vTPM does not fit well into the paradigm of cloud computing. Integration of trusted computing technology and cloud computing naturally requires to support the abstraction of Virtual Machine Group (VMG). vTPM is not sufficient for this purpose because each vTPM is bound to a single virtual machine, and the vTPMs bound to a set of virtual machines allocated to the same customer are deemed independent. This not only incurs inconvenience in supporting inter-vTPM cooperations, but also leads to security vulnerabilities. To bridge this gap, we propose TPM for cloud (TPMc), which is bound to a virtual machine group that may be allocated to the same customer or different customers who need the virtual machines to collaborate with each other. We give the detailed implementation of a TPMc prototype system, evaluate the performance of TPMc through extensive testing, and demonstrate that TPMc achieve a significant time-cost reduction, exceeding about 50% in seal and unseal operation, 60% in extend operation.

Objectives: Hardware virtualization is an evolving technology, because it has the potential to facilitate the consolidation of multiple workloads on a single physical host directly through cloud. We introduce the First-Last algorithm to minimize energy consumption and migration. Methods: The problem is treated as an instance of bin packing algorithm which focuses on saving energy. However the overloading of active hosts and the resultant migration are the drawbacks of this approach. This algorithm was tested on a set of heterogeneous virtual machines which have demands in both single and multi-dimension and came either statically or dynamically to the server. Findings: The virtual machine placement problem can be treated as a bin packing problem, wherein the physical machines/hosts are the bins and the virtual machines are the objects. It can then be solved through various approaches such as first fit, best fit, next fit et cetera. Using the first fit algorithm, the virtual machine is placed into the first active host which can accommodate it. If none of the active hosts satisfy the VM’s demand, then a new host is made active. In the best fit method, all the active hosts are checked first and then the host which would suffer minimum resource wastage upon VM placement is chosen. Next fit algorithm places the VM in the last /most recently activated host. While first fit is best suited for optimal performance, best fit keeps the least number of hosts active, thereby saving energy but at the cost of performance since migration overhead increases. Next fit approach would fail unless the virtual machines have been arranged in the increasing order of their demands. The firstlast algorithm that we introduce balances on energy consumption as well as performance by reducing migration. It also reduces the time complexity to a certain extent. Applications: The algorithm has a comparatively lower time complexity than the existing virtual machine placement algorithms. It also reduces the energy consumption as well as migration to a certain extent.

Load balancing is the core of virtual resource management and scheduling in cloud computing. For network applications, the cost of user would be greatly saved if load balancer could dynamically adjust cluster resources in accordance with the current applied load. The current load balancing products of cloud, such as Amazon's ELB, can be used to manage virtual machines in the cloud. The main drawbacks are still only supporting template-based deployment of new virtual machines, not supporting the trend prediction, failing to gain resources dynamically, and not sufficiently providing the elastic management of resources. Since the virtual machine for load balancing management in cloud computing can be dynamically applied and released, an algorithm of prediction-based elastic load balancing resource management (TeraScaler ELB) is presented to overcome the drawbacks. Experiments have shown that the required number of virtual machines change in compliance with the change of network load, thus TeraScaler ELB is able to dynamically adjust the processing capacity of back-end server cluster with the applied load. Besides it could make full use of the 'use on demand' feature of cloud computing, TeraScaler ELB leads to a better application of prediction based load balancing in cloud computing. It concludes that compared with the traditional elastic resource management algorithm, TeraScaler ELB is more reasonable for providing scalability and high availability.