Virtual Machine Failures Research Articles

Achieving Reliability in Cloud Computing by a Novel Hybrid Approach.

Cloud computing (CC) benefits and opportunities are among the fastest growing technologies in the computer industry. Cloud computing's challenges include resource allocation, security, quality of service, availability, privacy, data management, performance compatibility, and fault tolerance. Fault tolerance (FT) refers to a system's ability to continue performing its intended task in the presence of defects. Fault-tolerance challenges include heterogeneity and a lack of standards, the need for automation, cloud downtime reliability, consideration for recovery point objects, recovery time objects, and cloud workload. The proposed research includes machine learning (ML) algorithms such as naïve Bayes (NB), library support vector machine (LibSVM), multinomial logistic regression (MLR), sequential minimal optimization (SMO), K-nearest neighbor (KNN), and random forest (RF) as well as a fault-tolerance method known as delta-checkpointing to achieve higher accuracy, lesser fault prediction error, and reliability. Furthermore, the secondary data were collected from the homonymous, experimental high-performance computing (HPC) system at the Swiss Federal Institute of Technology (ETH), Zurich, and the primary data were generated using virtual machines (VMs) to select the best machine learning classifier. In this article, the secondary and primary data were divided into two split ratios of 80/20 and 70/30, respectively, and cross-validation (5-fold) was used to identify more accuracy and less prediction of faults in terms of true, false, repair, and failure of virtual machines. Secondary data results show that naïve Bayes performed exceptionally well on CPU-Mem mono and multi blocks, and sequential minimal optimization performed very well on HDD mono and multi blocks in terms of accuracy and fault prediction. In the case of greater accuracy and less fault prediction, primary data results revealed that random forest performed very well in terms of accuracy and fault prediction but not with good time complexity. Sequential minimal optimization has good time complexity with minor differences in random forest accuracy and fault prediction. We decided to modify sequential minimal optimization. Finally, the modified sequential minimal optimization (MSMO) algorithm with the fault-tolerance delta-checkpointing (D-CP) method is proposed to improve accuracy, fault prediction error, and reliability in cloud computing.

Improved Metaheuristic Based Failure Prediction with Migration Optimization in Cloud Environment

Cloud data centers consume high volume of energy for processing and switching the servers among different modes. Virtual Machine (VM) migration enhances the performance of cloud servers in terms of energy efficiency, internal failures and availability. On the other end, energy utilization can be minimized by decreasing the number of active, underutilized sources which conversely reduces the dependability of the system. In VM migration process, the VMs are migrated from underutilized physical resources to other resources to minimize energy utilization and optimize the operations. In this view, the current study develops an Improved Metaheuristic Based Failure Prediction with Virtual Machine Migration Optimization (IMFP-VMMO) model in cloud environment. The major intention of the proposed IMFP-VMMO model is to reduce energy utilization with maximum performance in terms of failure prediction. To accomplish this, IMFP-VMMO model employs Gradient Boosting Decision Tree (GBDT) classification model at initial stage for effectual prediction of VM failures. At the same time, VMs are optimally migrated using Quasi-Oppositional Artificial Fish Swarm Algorithm (QO-AFSA) which in turn reduces the energy consumption. The performance of the proposed IMFP-VMMO technique was validated and the results established the enhanced performance of the proposed model. The comparative study outcomes confirmed the better performance of the proposed IMFP-VMMO model over recent approaches.

A Broker-Based Task-Scheduling Mechanism Using Replication Approach for Cloud Systems

The reliability and availability of cloud systems have become major concerns of service providers, brokers, and end-users. Therefore, studying fault-tolerance mechanisms in cloud computing attracts intense attention in industry and academia. The task-scheduling mechanisms can improve the fault-tolerance level of cloud systems. A task-scheduling mechanism distributes tasks to a group of instances to be executed. Much work has been undertaken in this direction to improve the overall outcome of cloud computing, such as improving service quality and reducing power consumption. However, little work on task scheduling has studied the problem of lost tasks from the broker’s perspective. Task loss can happen due to virtual machine failures, server crashes, connection interruption, etc. The broker-based concept means that the backup task can be allocated by the broker on the same cloud service provider (CSP) or a different CSP to reduce costs, for example. This paper proposes a novel fault-tolerant mechanism that employs the primary backup (PB) model of task scheduling to address this issue. The proposed mechanism minimizes the impact of failure events by reducing the number of lost tasks. The mechanism is further improved to shorten the makespan time of submitted tasks in cloud systems. The experiments demonstrated that the proposed mechanism decreased the number of lost tasks by about 13%–15% compared with other mechanisms in the literature.

Modeling the Offloading Scheme With Unreliable VM-Related Service for MEC Networks

Multiaccess edge computing (MEC) allows terminal devices (TDs) with limited energy and computing capabilities to offload tasks to edge servers, which have more energy and computing capabilities to provide better service. In order to accommodate more users, an edge server can be divided into many virtual machines (VMs), but the interference among VMs could cause decreased service quality. At the same time, due to server hardware and software errors, there are random temporary failures regardless of whether a virtual machine is serving or not. This article uses a multidimensional Markov chain to establish a unified performance analysis model that jointly considers multiple types of terminals, different local service capabilities, different request arrival traffic, different virtual machine numbers, and different virtual machine failure rates. The statistical models of access blocking probability, resource availability, capacity, forced dropping probability, and retainability are presented. Furthermore, we analyze whether and how the local service should be performed when the virtual machine used fails and propose a backward offloading scheme. Analysis results show that some of quality is not monotonous. The model in this article provides a powerful and accurate analytical method for the preperformance simulation of actual engineering design.

Intelligent Fault-Tolerant Mechanism for Data Centers of Cloud Infrastructure

Fault tolerance in cloud computing is considered as one of the most vital issues to deliver reliable services. Checkpoint/restart is one of the methods used to enhance the reliability of the cloud services. However, many existing methods do not focus on virtual machine (VM) failure that occurs due to the higher response time of a node, byzantine fault, and performance fault, and existing methods also ignore the optimization during the recovery phase. This paper proposes a checkpoint/restart mechanism to enhance reliability of cloud services. Our work is threefold: (1) we design an algorithm to identify virtual machine failure due to several faults; (2) an algorithm to optimize the checkpoint interval time is designed; (3) lastly, the asynchronous checkpoint/restart with log-based recovery mechanism is used to restart the failed tasks. The valuation results obtained using a real-time dataset shows that the proposed model reduces power consumption and improves the performance with a better fault tolerance solution compared to the nonoptimization method.

OFP-TM: an online VM failure prediction and tolerance model towards high availability of cloud computing environments.

The indispensable collaboration of cloud computing in every digital service has raised its resource usage exponentially. The ever-growing demand of cloud resources evades service availability leading to critical challenges such as cloud outages, SLA violation, and excessive power consumption. Previous approaches have addressed this problem by utilizing multiple cloud platforms or running multiple replicas of a Virtual Machine (VM) resulting into high operational cost. This paper has addressed this alarming problem from a different perspective by proposing a novel mathbb {O}nline virtual machine mathbb {F}ailure mathbb {P}rediction and mathbb {T}olerance mathbb {M}odel (OFP-TM) with high availability awareness embedded in physical machines as well as virtual machines. The failure-prone VMs are estimated in real-time based on their future resource usage by developing an ensemble approach-based resource predictor. These VMs are assigned to a failure tolerance unit comprising of a resource provision matrix and Selection Box (S-Box) mechanism which triggers the migration of failure-prone VMs and handle any outage beforehand while maintaining the desired level of availability for cloud users. The proposed model is evaluated and compared against existing related approaches by simulating cloud environment and executing several experiments using a real-world workload Google Cluster dataset. Consequently, it has been concluded that OFP-TM improves availability and scales down the number of live VM migrations up to 33.5% and 83.3%, respectively, over without OFP-TM.

A Multi-Objective Genetic Algorithm-Based Resource Scheduling in Mobile Cloud Computing

Mobile cloud computing is an emerging technology in recent years. This technology reduces battery consumption and execution time by executing mobile applications in remote cloud server. The virtual machine (VM) load balancing among cloudlets in MCC improves the performance of application in terms of response time. Genetic algorithm (GA) is popular for providing optimal solution for load balancing problems. GA can perform well in both homogeneous and heterogeneous environments. In this paper, the authors consider multi-objective genetic algorithm for load balancing in MCC (MOGALMCC) environment. In MOGALMCC, they consider distance, bandwidth, memory, and cloudlet server load to find optimal cloudlet before scheduling VM in another cloudlet. The framework MOGALMCC aims to improve response time as well as minimizes VM failure rate. The experiment result shows that proposed model performed well by reducing execution time and task waiting time at server.

A New Adaptive Fault Tolerant Framework in the Cloud

The of robustness in cloud computing is becoming mainstream due to unprecedented growth in the cloud. In order to enhance cloud service reliability, various fault tolerance techniques have been suggested in the literature. Many cloud service reliability and availability issues are arising due to the failure of virtual machines (VMs). Therefore, this study contributes a new Threshold-Based Adaptive Fault Tolerance (TBAFT) approach in the cloud by amalgamating proactive and reactive approaches. Furthermore, in this paper, VM migration and replication are being utilized in the proactive approach, whereas retry is used as a reactive approach. Additionally, experimental validation is performed by comparing three benchmark fault tolerance approaches by incorporating the first fit and least full host selection policies. The obtained experimental results of the proposed approach are presented in terms of different metrics such as failure rate, throughput, migration, and execution time using CloudSim simulator.

Real-time and dynamic fault-tolerant scheduling for scientific workflows in clouds

Cloud computing has become a popular technology for executing scientific workflows. However, with a large number of hosts and virtual machines (VMs) being deployed, the cloud resource failures, such as the permanent failure of hosts (HPF), the transient failure of hosts (HTF), and the transient failure of VMs (VMTF), bring the service reliability problem. Therefore, fault tolerance for time-consuming scientific workflows is highly essential in the cloud. However, existing fault-tolerant (FT) approaches consider only one or two above failure types and easily neglect the others, especially for the HTF. This paper proposes a Real-time and dynamic Fault-tolerant Scheduling (ReadyFS) algorithm for scientific workflow execution in a cloud, which guarantees deadline constraints and improves resource utilization even in the presence of any resource failure. Specifically, we first introduce two FT mechanisms, i.e., the replication with delay execution (RDE) and the checkpointing with delay execution (CDE), to cope with HPF and VMTF, simultaneously. Additionally, the rescheduling (ReSC) is devised to tackle the HTF that affects the resource availability of the entire cloud datacenter. Then, the resource adjustment (RA) strategy, including the resource scaling-up (RS-Up) and the resource scaling-down (RS-Down), is used to adjust resource demands and improve resource utilization dynamically. Finally, the ReadyFS algorithm is presented to schedule real-time scientific workflows by combining all the above FT mechanisms with RA strategy. We conduct the performance evaluation with real-world scientific workflows and compare ReadyFS with five vertical comparison algorithms and three horizontal comparison algorithms. Simulation results confirm that ReadyFS is indeed able to guarantee the fault tolerance of scientific workflow execution and improve cloud resource utilization.

Design o f a Fault Tolerant Strategy f or Resource Scheduling i n Cloud Environment

Cloud computing supports the technological need of the industry supporting many other technologies. Also, the demand for computing power and storage by recent technologies is reasonably growing in a drastic way. Cloud computing, serving for these technologies are to be developed with advancements that lead to performance improvement both in support to the technologies like block-chain and big data. The allocation of cloud resources is an important strategy to be followed in a wiser manner to incorporate the needs of extra ordinary computing power. In this paper, an efficient resource allocation strategy (FTVMA) is introduced that involves the creation of effective virtual machines (VMs) and performs VM allocation in an efficient manner by considering the failure rates, previous history of failure of VM, execution efficiency as a part of effective scheduling. There exist many reasons for cloudlet failure in VMs. Some of them are overloading of VMs and non-availability of VMs. The introduced FTVMA algorithm considers the failure rate of the physical machine, load of virtual machines and the cost priority of the tasks in order to achieve Quality of Service (QoS) and Quality of Experience (QoE) of the user. The FTVMA methodology proposed in this paper works better for computation intensive VMs and is tested using CloudSim environment. The QoS metrics used to measure the performance of the proposed algorithm are Makespan and VM Utilization. The metric to measure QoE are Priority Miss Rate and Failure Rate. The proposed algorithm shows its improvement in terms of the QoS and QoE metrics. The results obtained are compared with the existing resource scheduling algorithms and it is inferred that the proposed algorithm performs better in terms of QoS and QoE.

A Strategy for Live Migration of Virtual Machines in a Cloud Federation

Migration of virtual machines (VMs) from one data center to another is essential to deal with scenarios such as load balancing, maintenance, power management, VM failures, etc. Application of VM migration is not only limited to a single cloud service provider, but also to inter-cloud architectures such as cloud federation. Migration in a federation is a complex operation as it encompasses many autonomous service providers with different standards, rules, and protocols. Migration can be carried out in two ways, first, live and, second, non-live. In case of live migration, a VM continues to execute at the source in contrast to non-live where its execution is suspended. In this paper, a framework for secure live migration of VMs in a cloud federation is proposed. Costs associated with it is analyzed for serial, parallel, and improved serial strategies. The proposed framework is simulated using a CloudSim simulator. Metrics considered for evaluation are communication overhead, migration time, and downtime. Further, power consumption for all strategies is calculated and compared. It is observed that downtime is least for parallel, whereas migration time is least for improved serial. The power consumption is highest for parallel and least for improved serial.

Evaluation of the impacts of failures and resource heterogeneity on the power consumption and performance of IaaS clouds

In this paper, we model the infrastructure of cloud computing systems to evaluate the power consumption and performance measures using stochastic activity networks (SANs). In the proposed model, servers run different numbers of virtual machines (VMs) and tasks are divided into two categories, namely low and high, demonstrating their priorities to run. Furthermore, both servers and VMs can fail during their operation. We consider the shutting down and the dynamic voltage and frequency scaling (DVFS) techniques for decreasing the power consumption in the proposed model. These techniques can also affect the performance of cloud systems. By evaluating the results obtained from the proposed SAN model in different scenarios, we conclude that the scenario in which servers run different numbers of VMs compared to scenarios at which servers run the same number of VMs is optimal in terms of both power and performance measures. Furthermore, the obtained results represent that failures do not much affect the power consumption, but the failure of servers, in comparison with the failure of VMs, has a great impact on the performance of the system under study. We also cross-validate the results obtained from the proposed analytical model, by applying the Mobius modeling tool, with the simulation results gained from the CloudSim framework.

A New Approach for VM Failure Prediction using Stochastic Model in Cloud

In the present study, a new approach is suggested for the prediction of virtual machine (VM) failure based on the time series stochastic model. Motivated by various proactive fault tolerance techniques in the cloud, prediction of VM failure is highly anticipated. However, precise prediction of VM failure is extremely challenging and having a substantial impact on the proactive fault tolerance system. In proactive fault management techniques, it is essential to forecast the VM failure in the cloud and to characterize the system behavior. To predict the VM failure, we need to monitor the execution of VMs in the clouds dynamic environment and gather their health-related information. Therefore, in this paper, VM failure prediction is presented through a failure predictor module which is designed using time series based autoregressive integrated moving average. The suggested approach is tested on non-stationary failure trace of VM. It is observed that the proposed approach provides the failure prediction accurately.

Fast and Reliable Restoration Method of Virtual Resources on OpenStack

We propose a fast and reliable restoration method of virtual resources on OpenStack when physical servers or virtual machines are down. Many providers have recently started cloud services, and the use of OpenStack, which is open source IaaS software, is increasing. When physical servers are down, there is a fail-over method using the high-availability cluster software such as Pacemaker to restore virtual resources. However, it takes a long time to restore all virtual resources. There is also a method for monitoring each virtual machine by using Ping or other methods and restoring a virtual machine when it is down. However, data may be destroyed due to the double mounts of virtual machines depending on the timing of failures because restoration methods of failed physical servers and virtual machines are independent. Therefore, we propose a fast and reliable restoration method with a uniform way for plural types virtual resources. In our method, Pacemaker only detects a physical server failure and notifies a failure to a virtual resource arrangement scheduler, then a virtual resource arrangement scheduler determines multiple physical servers to restore virtual resources and calls OpenStack APIs to rebuild. The virtual resource arrangement scheduler also detects virtual machine failures by using a Libvirt monitoring module and restores virtual machines without data loss by handling Pacemaker and Libvirt notifications uniformly. We implemented the proposed method and showed its effectiveness regarding fast restoration through performance measurements.

Handling Anti-Virtual Machine Techniques in Malicious Software

Malware analysis relies heavily on the use of virtual machines (VMs) for functionality and safety. There are subtle differences in operation between virtual and physical machines. Contemporary malware checks for these differences and changes its behavior when it detects a VM presence. These anti-VM techniques hinder malware analysis. Existing research approaches to uncover differences between VMs and physical machines use randomized testing, and thus cannot guarantee completeness. In this article, we propose a detect-and-hide approach, which systematically addresses anti-VM techniques in malware. First, we propose cardinal pill testing —a modification of red pill testing that aims to enumerate the differences between a given VM and a physical machine through carefully designed tests. Cardinal pill testing finds five times more pills by running 15 times fewer tests than red pill testing. We examine the causes of pills and find that, while the majority of them stem from the failure of VMs to follow CPU specifications, a small number stem from under-specification of certain instructions by the Intel manual. This leads to divergent implementations in different CPU and VM architectures. Cardinal pill testing successfully enumerates the differences that stem from the first cause. Finally, we propose VM Cloak —a WinDbg plug-in which hides the presence of VMs from malware. VM Cloak monitors each execute malware command, detects potential pills, and at runtime modifies the command’s outcomes to match those that a physical machine would generate. We implemented VM Cloak and verified that it successfully hides VM presence from malware.

Cloud Computing: A model Construct of Real-Time Monitoring for Big Dataset Analytics Using Apache Spark

The volume of data being collected, analyzed, and stored has exploded in recent years, in particular in relation to the activity on the cloud computing. While large-scale data processing, analysis, storage, and platform model such as cloud computing were previously and currently are increasingly. Today, the major challenge is it address how to monitor and control these massive amounts of data and perform analysis in real-time at scale. The traditional methods and model systems are unable to cope with these quantities of data in real-time. Here we present a new methodology for constructing a model for optimizing the performance of real-time monitoring of big datasets, which includes a machine learning algorithms and Apache Spark Streaming to accomplish fine-grained fault diagnosis and repair of big dataset. As a case study, we use the failure of Virtual Machines (VMs) to start-up. The methodology proposition ensures that the most sensible action is carried out during the procedure of fine-grained monitoring and generates the highest efficacy and cost-saving fault repair through three construction control steps: (I) data collection; (II) analysis engine and (III) decision engine. We found that running this novel methodology can save a considerate amount of time compared to the Hadoop model, without sacrificing the classification accuracy or optimization of performance. The accuracy of the proposed method (92.13%) is an improvement on traditional approaches.

Multiobjective Level-Wise Scientific Workflow Optimization in IaaS Public Cloud Environment

Cloud computing in the field of scientific applications such as scientific big data processing and big data analytics has become popular because of its service oriented model that provides a pool of abstracted, virtualized, dynamically scalable computing resources and services on demand over the Internet. However, resource selection to make the right choice of instances for a certain application of interest is a challenging problem for researchers. In addition, providing services with optimal performance at the lowest financial resource deployment cost based on users’ resource selection is quite challenging for cloud service providers. Consequently, it is necessary to develop an optimization system that can provide benefits to both users and service providers. In this paper, we conduct scientific workflow optimization on three perspectives: makespan minimization, virtual machine deployment cost minimization, and virtual machine failure minimization in the cloud infrastructure in a level-wise manner. Further, balanced task assignment to the virtual machine instances at each level of the workflow is also considered. Finally, system efficiency verification is conducted through evaluation of the results with different multiobjective optimization algorithms such as SPEA2 and NSGA-II.

Hybrid Fault Tolerant Scheme to Manage VM Failure in the Cloud

Objectives: To managing the Virtual Machine and improve the reliability and availability of the resources using Hybrid fault tolerance method. Methods/Statistical Analysis: Basically two main fault tolerant policies namely reactive fault tolerance and proactive fault tolerance. Reactive fault tolerant techniques attempts to reduce the effort required for recovery after a failure, whereas a proactive fault tolerant technique attempts to prevent the occurrence of a fault. The existing method aims to prevent the occurrence of VM failure using a proactive fault tolerant technique which is a static method. But it provides no solution for VM failure, which might occur while the application is still running. The proposed paper ensures reliability of VM by considering both the reactive and proactive techniques in hybrid ways of tolerating the VM failure. In the proposed system, two types of VM failure, insufficient storage and variable capacity change are detected. These failures are tolerated using a self healing mechanism implemented at the host level. Result: The proposed scheme is implemented using Cloudsim simulator, in order to highlight the performance of how VM can be managed. Applications: Virtual Machine failure can be managed and improve the reliability and availability of resources using hybrid fault tolerance scheme.

Reliable self-deployment of distributed cloud applications

Cloud applications consist of a set of interconnected software elements distributed over several virtual machines, themselves hosted on remote physical servers. Most existing solutions for deploying such applications require human intervention to configure parts of the system, do not conform to functional dependencies among elements that must be respected when starting them, and do not handle virtual machine failures that can occur when deploying an application. This paper presents a self-deployment protocol that was designed to automatically configure a set of software elements to be deployed on different virtual machines. This protocol works in a decentralized way, that is, there is no need for a centralized server. It also starts the software elements in a certain order, respecting important architectural invariants. This protocol supports virtual machine and network failures and always succeeds in deploying an application when faced with a finite number of failures. Designing such highly parallel management protocols is difficult; therefore, formal modeling techniques and verification tools were used for validation purposes. The protocol was implemented in Java and was used to deploy industrial applications. Copyright © 2016 John Wiley & Sons, Ltd.

Formal design of dynamic reconfiguration protocol for cloud applications

Cloud applications are complex applications composed of a set of interconnected software components running on different virtual machines, hosted on remote physical servers. Deploying and reconfiguring this kind of applications are very complicated tasks especially when one or multiple virtual machines fail when achieving these tasks. Hence, there is a need for protocols that can dynamically reconfigure and manage running distributed applications. In this article, we present a novel protocol, which aims at reconfiguring cloud applications. This protocol is able to ensure communication between virtual machines and resolve dependencies by exchanging messages, (dis)connecting, and starting/stopping components in a specific order. The interaction between machines is assured via a publish-subscribe messaging system. Each machine reconfigures itself in a decentralized way. The protocol supports virtual machine failures, and the reconfiguration always terminates successfully even in the presence of a finite number of failures. Due to the high degree of parallelism inherent to these applications, the protocol was specified using the LNT value-passing process algebra and verified using the model checking tools available in the CADP toolbox. The use of formal specification languages and tools helped to detect several bugs and to improve the protocol.