Live Migration Of Virtual Machines Research Articles

Novel probabilistic resource migration algorithm for cross-cloud live migration of virtual machines in public cloud

In cloud computing environment, cross-cloud live migration of virtual machines (VMs) is a major concern in these days. Cloud computing provides the users with huge, versatile and on-demand access to a bulk of customizable and configurable registered physical devices or things. It helps organizations or enterprises to share data efficiently by privately owned cloud or by the third-party servers. This type of sharing of bulky data through cloud is more efficient and reliable. In an enterprise environment, one of the essential capabilities of cloud infrastructure is VM migration. VM live migration basically involves the transference of instances that includes the operating system, runtime memory pages and active CPU states from source hub to the destination hub. In this paper, we have discussed on resource allocation algorithm which performs better in utilization of CPU, time and memory. Our proposed algorithm deals with the effective utilization of unoccupied memory, and we have also measured VM memory stack flow of total memory for cloud computing architecture.

Energy-Efficient Virtual Machine Scheduling across Cloudlets in Wireless Metropolitan Area Networks

Nowadays, with the development of wireless communication, people are relying on mobile devices heavily due to various application deployment and plentiful service experience of customers in wireless metropolitan area network (WMAN). The computing resources of the mobile devices are limited as they are restricted with physical size, battery capacity, etc. In order to release the resource limitation, cloudlet, an effective paradigm, is introduced to host the computing tasks offloaded from the mobile devices. Compared with the cloud computing, the cloudlets are deployed closer to provide customers with fewer task offloading delay. Currently, the offloaded task scale is increasing, which generates large quantities of energy exposure for task implementation among cloudlets. Taking advantage of live virtual machine (VM) migration technology, the energy consumption of cloudlets could be definitely reduced. But such migration across cloudlets also decreases the implementation performance of the tasks. Therefore, it is still a challenge to jointly optimize the execution performance and the energy consumption for cloudlet management in WMAN. In this paper, a novel virtual machine (VM) scheduling method is proposed to balance the implantation time and the energy consumption to cope with the above challenge. Specifically, a collection of available migration polices are obtained through heuristically searching of destination cloudlets for the running computing tasks. Then, simple additive weighting (SAW) and multiple criteria decision making (MCDM) techniques are leveraged to select the optimal VM scheduling strategy across cloudlets in WMAN. Finally, experimental results and evaluations validate our proposed method is both effective and feasible.

Performance evaluation of live virtual machine migration in SDN-enabled cloud data centers

In Software-Defined Networking (SDN) enabled cloud data centers, live VM migration is a key technology to facilitate the resource management and fault tolerance. Despite many research focus on the network-aware live migration of VMs in cloud computing, some parameters that affect live migration performance are neglected to a large extent. Furthermore, while SDN provides more traffic routing flexibility, the latencies within the SDN directly affect the live migration performance. In this paper, we pinpoint the parameters from both system and network aspects affecting the performance of live migration in the environment with OpenStack platform, such as the static adjustment algorithm of live migration, the performance comparison between the parallel and the sequential migration, and the impact of SDN dynamic flow scheduling update rate on TCP/IP protocol. From the QoS view, we evaluate the pattern of client and server response time during the pre-copy, hybrid post-copy, and auto-convergence based migration.

Bio-Inspired Heuristics for VM Consolidation in Cloud Data Centers

In infrastructure-as-a-service environments, Cloud data centers employ virtualization technology to host various applications in virtual machines (VMs) and enable application isolation on shared physical resources. Additionally, live VM migration has been adopted to perform load balancing by moving VMs across distinct hosts. However, co-located VMs that show significant positive correlations on their CPU utilization patterns are at a higher risk of triggering overloading events and incurring performance degradation, even when their host is operating below its critical limits. To address this problem, a VM consolidation mechanism inspired by host-switching behaviors in symbiotic associates is proposed in this paper. In the proposed mechanism, hosts and VMs in Cloud data centers represent symbionts in an ecosystem. Two heuristic functions, inspired by host susceptibility and symbiotic coefficient among symbionts, are proposed to yield better resource utilization via VM consolidations. Experiment results demonstrate that the proposed mechanism can achieve reductions in both energy consumption and service level agreement violations of Cloud data centers.

An energy-efficient cloud system with novel dynamic resource allocation methods

In the last decade, cloud computing has brought enormous changes to people’s lives. Cloud computing gives a client-driven computational model. In this case, the pay-per-use model enables a flexible method of deploying and sharing distributed services and resources. Furthermore, computational resources are progressively reassembled and insulated to give various types of services. Current solutions on how to efficiently utilize energy consumption are inadequate. Thus, we propose two methods to tackle the problem, which are the dynamic resource allocation method and the energy saving method, respectively. In this work, we firstly deployed an infrastructure platform based on OpenStack to actualize the proposed approach with live migration of virtual machines (VMs). Secondly, we allocated the dynamic resources and proposed the energy saving algorithm. Thirdly, we monitor the power distribution unit status to record the energy consumption of the system. Finally, in the experiments, we found that the proposed algorithms cannot just accomplish the proficient use of VMs’ resources but also thrift the energy usage of physical machines. Our experimental results demonstrate about 39.89% of energy is saved from 20 VCPUs of 20 GB memory.

DCnet: a data centre network architecture that supports live VM migration

Cloud computing and storage have become an integral part of online Internet services. Over the past decade, enterprises have steadily moved their applications and storage to the cloud paradigm. Server virtualisation enables cloud providers to utilise computational resources efficiently by employing virtual machines (VMs) and containers. As more enterprise applications are hosted by VMs in the cloud, a significant amount of network communication occurs among servers within a data centre, increasing so-called ‘ east-west communication’ . In addition to scaling to accommodate more east-west traffic, future data centre networks must support efficient VM migration, which is used to optimise power consumption. This article proposes a new data centre network architecture called DCnet . DCnet changes addressing and routing at layers 2 and 3 completely to increase throughput and support live VM migration throughout an organisation, including across multiple data centres owned by an organisation that spans large geographic areas. DCnet retains compatibility with existing hardware by using the same frame format as standard protocols, specifically Ethernet and Internet protocol (IP). Furthermore, DCnet does not require any changes to applications and host operating systems or libraries. In addition to presenting the architecture, the article describes a test bed used to assess DCnet and reports experimental measurements.

Adaptive deduplication of virtual machine images using AKKA stream to accelerate live migration process in cloud environment

Cloud Computing is a paradigm which provides resources to users from its pool based on demand to satisfy their requirements. During this process, many servers are overloaded and underloaded in the cloud environment. Thus, power consumption and load balancing are the major problems and are resolved by live virtual machine (VM) migration. Load balancing is addressed by moving virtual machines from overloaded host to under loaded host and from under loaded host to any other host which is not overloaded called VM migration. If this process is done without power off (Live) the virtual machines then it is called live VM migration. By this process, the issue of power consumption by physical hosts is also resolved. Migrating virtual machines involves virtualized components like storage disks, memory, CPU and networking, the entire state of VM is captured as a collection of data files. These data files are virtual disk files, configuration files, and log files. The virtual disk files take larger memory and take more migration time and hence the downtime. These disk files include many zero pages, similar and redundant pages. Many techniques such as compression, deduplication is used reduce the size of VM disk image file. Compression techniques are not widely used, due to the disadvantage of compression ratio and decompression time. Many researchers hence used deduplication techniques for reducing the VM disk image file in the live migration process. The significance of the research work is to design an adaptive deduplication mechanism for reducing VM disk image file size by performing fixed length and variable length block-level deduplication processes. The Rabin-Karp rolling hash algorithm is used in variable length block-level deduplication. Akka stream is used for streaming the VM disk image files as it is the bulk volume of live data transfer. To reduce the time of the deduplication process, many researchers used multithreading and multi-core technologies. We use multithreading in Akka framework to run the deduplication process concurrently without OutofMemory errors. The experiment results show that we achieved a maximum of 83% overall reduction in image storage space and 89.76% reduction in total migration time are achieved by adaptive deduplication method. 3% improvement in deduplication rate when compared with the existing image management system. The results are significant because when we apply this in the storage of data centres, there are much space savings. The reduction in size is dependent on the dataset was taken and the applications running on the VM.

Efficient live virtual machine migration for memory write-intensive workloads

Live migration of virtual machines (VMs) is an indispensable feature in cloud computing systems due to the use of load balancing, system maintenance, power management, etc. However, it is usually an intractable problem to migrate the VMs running memory write-intensive workloads, because they rapidly generate plenty of dirty pages, which need to be resent many times. Migration failures or compromises with large downtimes usually result in severe consequences such as the violation of service level agreements, abortion of TCP connections, etc. To solve this challenging problem, we first demonstrate a novel observation: during the pre-copy migrations, some workloads have a large portion of “fake dirty” pages, which are unnecessary and wasteful to be resent. After exploring how they are generated, we propose to leverage the secure hash method to avoid resending them. Besides, to guarantee the successful accomplishment of migration, we further propose the intelligent hybrid migration. Our scheme leverages heuristic and automatically switches from pre-copy to post-copy at the near-optimal moment to obtain a short post-copy duration, thus alleviating post-copy’s inherent weaknesses. Evaluations show that our proposed scheme gets a significant performance improvement for VM migration. The workloads which fail to be migrated with pre-copy now accomplish migration quickly, with a total migration time from 27 s to 98 s. Besides, the intelligent hybrid migration remarkably shortens the post-copy duration by an extent from 43% to 60%, compared with the traditional hybrid method.

Live Placement of Interdependent Virtual Machines to Optimize Cloud Service Profits and Penalties on SLAs

This paper aims to optimize cloud services' net profits and penalties with live placement of interdependent virtual machines (VMs). This optimization is a complex task as it is difficult to achieve a successful compromise between penalties and net profits on service level contracts. This paper studies this optimization problem to minimize services' penalties and maximizing net profits while achieving live migrations of interdependent VMs. This VM's live placement optimization problem is a NP-hard problem with exponential running time. A mathematical model was designed and approximations were conducted with an efficient PCH/PCH' heuristic. This Mixed Integer Non-Linear programming (MNLP) formulation and heuristic for cloud services was tested where the overall services' penalty needs to be minimized, overall net profits have to be maximized, and where efficient live migrations of VMs is a concern. Simulation results show how cloud providers may live place VMs. Finally, our results show that a PCH/PCH' heuristic: (i) finds better solutions than the existing machines' configuration of Google traces; (ii) is suitable for large-sized instances of cloud services; (iii) performs better than FF, FFD, and CPLEX in terms of overall penalties and net profits; and (iv) runs in less than six minutes over the last day's data.

Mobility-aware task delegation model in mobile cloud computing

Mobile devices frequently move with different velocities. Task delegation to remote cloud servers becomes critical when the requesting device changes its location. The communication with the remote cloud server might be lost, and the user does not receive the result. To solve this problem, two analytical models are provided in this article. In the first solution, we propose an approach where the mobile device uses remote cloud servers for executing task. When the user changes location, the mobile device looses connection with the cloud instance. Once the result is ready, the cloud server pushes the result back to the device via push notification message when the device reconnects with the network. However, the mobile device can also get the result by serializing session information. In the second solution, the mobile device offloads code into cloudlets. When the mobile device changes location, virtual machine live migration happens. The present state of the instance is transferred from the previous cloudlet to the new cloudlet, where the offloading process resumes execution. It is observed that the proposed task delegation and code-offloading models reduce the power consumptions, respectively, by 30–63% and 61–78% approximately than the existing mobility-aware approach. Experimental results are obtained using mobile device with various velocities inside and outside the university building.

Towards Green Cloud Computing an Algorithmic Approach for Energy Minimization in Cloud Data Centers

The article presents an efficient energy optimization framework based on dynamic resource scheduling for VM migration in cloud data centers. This increasing number of cloud data centers all over the world are consuming a vast amount of power and thus, exhaling a huge amount of CO2 that has a strong negative impact on the environment. Therefore, implementing Green cloud computing by efficient power reduction is a momentous research area. Live Virtual Machine (VM) migration, and server consolidation technology along with appropriate resource allocation of users' tasks, is particularly useful for reducing power consumption in cloud data centers. In this article, the authors propose algorithms which mainly consider live VM migration techniques for power reduction named “Power_reduction” and “VM_migration.” Moreover, the authors implement dynamic scheduling of servers based on sequential search, random search, and a maximum fairness search for convenient allocation and higher utilization of resources. The authors perform simulation work using CloudSim and the Cloudera simulator to evaluate the performance of the proposed algorithms. Results show that the proposed approaches achieve around 30% energy savings than the existing algorithms.

Performance of Memory Virtualization Using Global Memory Resource Balancing

Virtualization has become a universal generalization layer in contemporary data centers. By multiplexing hardware resources into multiple virtual machines and facilitating several operating systems to run on the same physical platform at the same time, it can effectively decrease power consumption and building size or improve security by isolating virtual machines. In a virtualized system, memory resource supervision acts as a decisive task in achieving high resource employment and performance. Insufficient memory allocation to a virtual machine will degrade its performance drastically. On the contrasting, over allocation reasons ravage of memory resources. In the meantime, a virtual machine's memory stipulates may differ drastically. As a consequence, effective memory resource management calls for a dynamic memory balancer, which, preferably, can alter memory allocation in a timely mode for each virtual machine-based on their present memory stipulate and therefore realize the preeminent memory utilization and the best possible overall performance. Migrating operating system instances across discrete physical hosts is a helpful tool for administrators of data centers and clusters: It permits a clean separation among hardware and software, and make easy fault management. In order to approximate the memory, the stipulate of each virtual machine and to adjudicate probable memory resource disagreement, an extensively planned approach is to build an Least Recently Used based miss ratio curve which provides not only the current working set size but also the correlation between performance and the target memory allocation size. In this paper, the authors initially present a low overhead LRU-based memory demand tracking scheme, which includes three orthogonal optimizations: AVL based Least Recently Used association, dynamic hot set sizing. This assessment outcome confirms that, for the complete SPEC CPU 2006 benchmark set, subsequent to pertaining the 3 optimizing techniques, the mean overhead of MRC construction are lowered from 173% to only 2%. Based on current WSS, the authors then predict its trend in the near future and take different tactics for different forecast results. When there is an adequate amount of physical memory on the host, it locally balances its memory resource for the VMs. Once the local memory resource is insufficient and the memory pressure is predicted to sustain for a sufficiently long time, VM live migration, is used to move one or more VMs from the hot host to other host(s). Finally, for transient memory pressure, a remote cache is used to alleviate the temporary performance penalty. These experimental results show that this design achieves 49% center-wide speedup.

Online User Distribution-Aware Virtual Machine Re-Deployment and Live Migration in SDN-Based Data Centers

Virtual machine (VM) re-deployment and migration has been proven to be a key technique for cloud data centers to implement resource optimization and load balance. Also, live VM migration aims to guarantee the continuity of the existing data flows. Though VM management has been well studied, it has so far mostly focused on optimizing resource utilization, while the important issue of users' quality of experience has been ignored. In mobile cloud computing, user distribution changes dynamically over time and it can significantly affect both the service latency and network resource utilization. In the traditional network, VM migration may cause a flash crowd of flow changing and a long service downtime due to VM's IP address changing. Software-defined networking is an emerging paradigm to logically centralize the network control plane and automate the configuration of individual network elements, which can be ubiquitously deployed in data centers and serve as an effective means for flow handling. In this paper, we design a user distribution-aware virtual machine re-deployment mechanism and propose a traffic-redirection virtual machine migration scheme to keep the active flows from being interrupted. We further provide simulations to show the effectiveness and superiority of the proposed approaches over existing schemes.

A prediction‐based and power‐aware virtual machine allocation algorithm in three‐tier cloud data centers

SummaryWith the increasing popularity of cloud computing services, the more number of cloud data centers are constructed over the globe. This makes the power consumption of cloud data center elements as a big challenge. Hereby, several software and hardware approaches have been proposed to handle this issue. However, this problem has not been optimally solved yet. In this paper, we propose an online cloud resource management with live migration of virtual machines (VMs) to reduce power consumption. To do so, a prediction‐based and power‐aware virtual machine allocation algorithm is proposed. Also, we present a three‐tier framework for energy‐efficient resource management in cloud data centers. Experimental results indicate that the proposed solution reduces the power consumption; at the same time, service‐level agreement violation (SLAV) is also improved.

Energy-efficient virtual machine selection based on resource ranking and utilization factor approach in cloud computing for IoT

IoT leads to abrupt variations producing an immense number of data streams for storage, which is a considerable task in the heterogeneous cloud computing environment. Extant techniques consider task deadlines for virtual machine (VM) allocation and migration. This creates a resource famine leading to haphazard and numerous VM migrations, high energy consumption and unbalanced resource utilization. To solve this issue, an energy-efficient resource ranking and utilization factor-based virtual machine selection (ERVS) approach is proposed. ERVS encompasses the resource requirement rate for task classification, comprehensive resource balance ranking, processing element cost and the resource utilization square model for migration. It evaluates overloaded and underloaded hosts and types of VM by predicting CPU utilization rate and energy consumption. Based on this, tasks are sorted and VMs are optimally assigned, which enhances the resource utilization rate, reducing the number of live VM migrations. The experiments evaluate the ability of the proposed approach to diminish energy consumption without violation of service level agreements.

Network-aware energy saving multi-objective optimization in virtualized data centers

With the current growth of data centers, improving energy saving is becoming more important to cloud service providers. The data centers architectural design and the advancement of virtualization technologies can be exploited for energy saving. In this paper, we studied the energy saving problem in data centers using virtual machines placement and live migration taking to account the status of the network links load. The problem was formulated as multi-objective integer linear program, which solvable by CPLEX, to minimize the energy consumed by the servers and minimize the time to migrate virtual machines. To overcome CPLEX high computation, a heuristic algorithm is introduced to provide practical and efficient virtual machines placement while minimizing their migration overhead to the network. The heuristic is evaluated in terms of energy consumed and performance using a real data center testbed that is stressed by running Hadoop Hibench benchmarks. The results where compared to the ones obtained by distributed resource scheduler (DRS) and the base case. The results show that the heuristic algorithm can save up to 30% of the server’s energy. For scalability and validity of optimality, the results of the heuristic were compared to the ones provided by CPLEX where the gap difference was less than 7%.

ABDM : Agent Based Live Migration of Virtual Machines in Cloud Computing For Multimedia Data

Migration time is one of the metric to measure the performance of the algorithm for live migration. In this paper we have introduced a new parameter for live migration of virtual machines (VM) called the ‘Exit Time’ which is defined as the time to eject the state of one or more VMs from the source node. Exit Time defines how rapidly the VM can be taken out from the source node and its resources are freed for reallocating other tasks. We present an Agent Based Live Migration which disconnects the source node from the destination node during migration to reduce the exit time if the destination is slow. The source distributes the memory of VMs to multiple intermediate nodes organized by a middleware. Simultaneously, the destination collects and merges the VMs’ memory from the intermediate nodes. Thus exit from the source node is no longer resisted by the receiving speed of the destination. We support simultaneous live exit of multiple VMs and our ABDM implementation in the CloudSim platform reduces the exit time by a considerable amount against the traditional pre-copy and post-copy migration at the same time keeping the total migration time when the destination node is sluggish than the source

A Versioning Approach to VM Live Migration

In the context of virtual machines live migration, two strategies called “pre-copy” and “post-copy” have already been presented; but each of these strategies works well only in some circumstances. In this paper, we have a brief presentation of QAVNS and then introduce a new approach which is based on the concept of object, assigning QAVNS-scheme-revision number, and observing the changes. In this approach, the total virtual machine memory is considered as a QAVNS informational object that is constantly changing. In this regard, we have defined some criteria and presented an algorithm by which the hypervisor can detect the current behavior of the virtual machine, and automatically select the virtual machine migration strategy from the two pre-copy and post-copy options. We evaluated the implementation & simulation platforms considering the state of the art available technologies, mostly CloudSim, SimGrid and DartCSim+. Formal analysis shows that applying the proposed scheme and the proposed algorithm can significantly improve the live migration process of virtual machines.

CBase: Fast Virtual Machine storage data migration with a new data center structure

Live Virtual Machine (VM) migration within a data center is an important technology for cloud management, and has brought many benefits to both cloud providers and users. With the development of cloud computing, across-data-center VM migration is also desired. Normally, there is no shared storage system between data centers, hence the storage data (disk image) of a VM will be migrated to the destination data center as well. However, the slow network speed of the Internet and the comparatively large size of VM disk image make VM storage data migration become a bottleneck for live VM migration across data centers. In this paper, based on a detailed analysis of VM deployment models and the nature of VM image data, we design and implement a new migration system, called CBase. The key concept of CBase is a newly introduced central base image repository for reliable and efficient data sharing between VMs and data centers. With this central repository, further performance optimizations to VM storage data migration are made possible. Two migration mechanisms (data deduplication and Peer-to-Peer (P2P) file sharing) are utilized to accelerate base image migration, and a strategy is designed to elevate the synchronization of newly-written disk blocks. The results from an extensive experiment show that CBase significantly outperforms existing migration mechanisms under different conditions regarding total migration time and total network traffic. In particular, CBase with data deduplication is better than P2P file sharing for base image migration in our experimental environment.

A Review on Different Types of Live VM Migration Methods with Proposed Pre-Copy Approach

Cloud computing is being considered as the future architecture of IT world. Virtualization creates logical resources from physical resources which are allocated with flexibility to applications. Server virtualization is a technique for the division of the physical machine into many Virtual Machines; every Virtual Machine has the capacity of applications execution similar to physical machine. The capability of Virtual Machine migration i.e. dynamic movement of Virtual Machines between physical machines is achieved by virtualization. Migration techniques differ w.r.t order of state transfer. Pre-copy migration method transfers all pages of memory from source to destination while Virtual Machine is executing on source. Post-copy migration is transfer of memory. content after the transfer of process state. Specially, post-copy migration, first copied the process states to the destination machine. Total time of migration, Total pages transferred and Downtime are important parameters considered during live Virtual Machine migration. Many improved live pre copy Virtual Machine migration techniques tries to decrease all the three above mentioned parameters. Proposed approach also tries to minimize all the three performance parameters.