Optimizing Virtual Machine Resource Placement on Multi-Socket Platforms

Measuring and analyzing code authorship in 1 + 118 open source projects

Recently, virtualization technology has revived and become the key support of the clouds. KVM (Kernel virtual machine) based on Linux kernel-level achieves its popularity however with security risks. Therefore, the virtual environment and vulnerability detecting of KVM is of theoretical significance and great application value. However, there are few reports on virtual environment and vulnerability detection for KVM. This paper introduces a virtual environment and vulnerability method which includes the CPU cycle based detection algorithms, the internet time detection algorithms and multi-thread counter detection algorithms, etc.. In the vulnerability detection of KVM, denial of service attack is simulated and its detection method is proposed. Function and performance tests are also introduced.

A common strategy to manage memory resources of VMs under changing workloads is Dynamic VM memory management via hotplug. However, in order to estimate the VM working set size, most researchers are utilizing approaches that rely on kernel instrumentation but this most often results to high runtime overhead. This will result in system administrate to exercise a tradeoff between the estimate accuracy and system performance. The novelty of this work is to present a light weight accurate and transparent prediction algorithm for re-balancing memory resources among VMs. Experiments result attained on Dacapo and SPECjvm2008 from renowned benchmarks shows that with only 4% performance overhead our proposed method is capable of accurately adjusting virtual machine memory size on its real time requirements, and improve application performance in the virtual machine more than 10% better when virtual machine has 2 CPUs and 20% better when it has 4 CPUs. In case there is no free memory available our proposed method will try first to use host's free memory but if there is no free memory, it starts memory over-commitment for a fixed duration to help the requested VM complete its task.

Cloud Computing is the distributed architecture in which scalable resources are used to provide computing services to user using virtualization concept. Cloud requires virtual machine for accessing any guest operating system at local desktop, this is handled by software interface such as Virtual Machine Monitor (VMM) or hypervisor. This paper reveals not only its features architecture, working but also comparative analysis of Kernel Virtual Machine (KVM) and Xen using various benchmarking tools. Paper also explains test analysis of measuring CPU performance, network speed and disk access. From this analysis which hypervisor has to be used for the implementation of cloud computing system for getting better performance is decided.

The article analyzes the advantages of mobile cloud technologies and problems emerging during the use of those. The network infrastructure created based on cloudlets at the second level of mobile cloud computing with hierarchical structure is analyzed. At the same time, the article explores the issues of satisfaction of demand of mobile equipment for computing and memory resources by using these technologies. The article presents one solution for the allocation of mobile user requests in virtual machines created in cloudlets located near base stations of wireless metropolitan area networks (WMAN) in a balanced way by considering the technical capacity of those. Alongside, the article considers the solution of user problem during designated time and the issue of determining virtual machines satisfying other requirements. For this purpose, different characteristics of the stated problem, virtual machines, as well as communication channels between a user and virtual machines are considered. By using possible values determining the importance of cloudlets, conditions for loading software applications of a user to a virtual machine are explored and an appropriate method is proposed.

Future in-car applications such as autonomous driving call for the use of multicore system-on-chips (SoCs) together with powerful hardware accelerators. Hypervisor technologies offer the opportunity to partition such platforms and efficiently orchestrate the use of the shared hardware accelerators. Moreover, this allows one to consolidate existing and brand new applications in a single platform and thereby preserve past investments. On the other hand, the use of hypervisor-technology also addresses the need for shielding mission-critical software from uncritical software applications co-located on the same platform. Integrating applications with different safety and security needs into a single platform raises manifold design questions. We, at Elektrobit, believe, that a microkernel-based Hypervisor solution can be a mean to establish spatial and temporal isolation of software stacks and therefore is vital for deploying high-performance SoCs, not only inside vehicles, but also in other domains where software systems have to be shielded from another. As part of Elektrobits contribution to the EPI project, a hypervisor demonstration stack with different virtual machines will be deployed on a performance-centric multicore platform. In this paper, we present our system layout and discuss its key design features.

Cloud platforms often take advantage of virtualization technology and make their actual hosts virtualized. As network attack events occur frequently, providing system security in a virtualized environment is the focus of this study. We have designed and implemented a lightweight network-based intrusion prevention system (IPS) named VMM-IPS for the virtual machine (VM) execution environment. To ensure the system safety of VMs and the host system at the same time, VMM-IPS is operated in the Linux kernel of the host system and co-located with the Kernel-based Virtual Machine that turns Linux kernel into a hypervisor. As packets enter the system, no matter destined to VMs or passing through the host, they are detected by VMM-IPS. Unlike user-level IPS that needs switching protection domain and copying packets to user buffer for inspection, VMM-IPS is more efficient because of the capability to perform in-place packet inspection. It adopts signature-based detection and is implemented with the multiple-pattern search algorithm AC-BM for efficient string matching. Besides, VMM-IPS can protect the system against attacks using packet splitting and reassembly to evade introduction detection system (IDS). The experimental results demonstrate VMM-IPS can achieve system safety effectively and efficiently.

The high use of system calls in many applications has motivated research works focusing on reducing the overhead caused by these calls. In this work we implement three types of compound system calls, which are evaluated experimentally by taking into account their execution time in systems with 1 to 8 cores. We also investigate the system calls running in both physical and virtual machines. All proposed compound system calls present performance gains when compared to their conventional counterparts. The results indicate reduction in the execution time up to 58 percent. The dataset obtained through controlled experiments is analyzed using ANOVA and Tukey tests. We also verify that the highest improvements occur in the non-virtualized environment.

Linux is an open source operating system that uses the Linux kernel. The security of Linux makes its users rarely use antivirus to prevent the threat of ransomware attacks. The purpose of this study is to assess the security of the Linux operating system that uses Wine as a compatibility layer against ransomware attacks. Then to determine whether ransomware has a higher impact than malware samples in previous studies or not. The methods used on the 30 samples are static analysis to identify samples, and dynamic analysis to determine the implications. From the 30 ransomware samples that can run on the Linux operating system, the results obtained show that 80% affect the registry, 50% affect the file system, 50% affect the service, 70% affect the Network and 60% affect the Process. While overall there are 30% of samples that affect the five existing parameters. The results show that ransomware has relatively high implications for existing parameters, inversely proportional to previous research.

Heterogeneous computing leverages more than one kind of processors to boost the performance of user-space applications with the heterogeneous programming languages, e.g., OpenCL. While some works have been done to accelerate the computations required by Linux kernel software, they are either application-specific solutions or tightly coupled with the certain computing platforms and are not able to support the general-purpose in-kernel accelerations using different types of processors. In this article, the general-purpose software framework called Kernel acceleration with OpenCL (KOCL), is proposed to tackle the problem. KOCL exposes a set of the high-level programming interfaces for the Linux kernel module developers to offload compute-intensive tasks on different hardware accelerators without managing and coordinating the platform-specific computing and memory resources. The simplified programming efforts are achieved by the developed platform management and memory models, which provide a systematic means of managing the heterogeneous hardware resources. In addition, the one- and zero-copy data-buffering schemes are offered by KOCL, so that the offloaded tasks deliver high performance on the platforms with different memory architectures. We have developed the prototype system to accelerate the Network-Attached Storage server applications. Significant performance improvements are achieved with the three different types of accelerators, i.e., the multicore processor, the integrated GPU, and the discrete GPU, respectively. We believe that KOCL is useful for the design of embedded appliances to evaluate the performance of design alternatives.

In general, operating systems (OSs) are designed to mediate access to device hardware by applications. They process different kinds of system calls using an indiscriminate kernel with the same configuration. Applications in cloud computing platforms are constructed from service components. Each of the service components is assigned separately to an individual virtual machine (VM), which leads to homogeneous system calls on each VM. In addition, the requirements for kernel function and configuration of system parameters from different VMs are different. Therefore, the suit-to-all design incurs an unnecessary performance overhead and restricts the OS’s processing capacity in cloud computing. In this paper, we propose an adaptive model for cloud computing to resolve the conflict between generality and performance. Our model adaptively specializes the OS of a VM according to the resource-consuming characteristics of workloads on the VM. We implement a prototype of the adaptive model, vSpec. There are five classes of VM: CPU-intensive, memory-intensive, I/O-intensive, networkintensive and compound, according to the resource-consuming characteristics of the workloads running on the VMs. vSpec specializes the OS of a VM according to the VM class. We perform comprehensive experiments to evaluate the effectiveness of vSpec on benchmarks and real-world applications.

Cloud computing has emerged as a new and alternative approach for providing computing services. Customers acquire and release resources by requesting and returning virtual machines to the cloud. Different service models and pricing schemes are offered by cloud service providers. This can make it difficult for customers to compare cloud services and select an appropriate solution. Cloud Infrastructure-as-a-Service vendors offer a t-shirt approach for Virtual Machines (VMs) on demand. Customers can select from a set of fixed size VMs and vary the number of VMs as their demands change. Private clouds often offer another alternative, called time-sharing, where the capacity of each VM is permitted to change dynamically. With this approach each virtual machine is allocated a dynamic amount of CPU and memory resources over time to better utilize available resources. We present a tool that can help customers make informed decisions about which approach works most efficiently for their workloads in aggregate and for each workload separately. A case study using data from an enterprise customer with 312 workloads demonstrates the use of the tool. It shows that for the given set of workloads the t-shirt model requires almost twice the number of physical servers as the time share model. The costs for such infrastructure must ultimately be passed on to the customer in terms of monetary costs or performance risks. We conclude that private and public clouds should consider offering both resource sharing models to meet the needs of customers.

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.

Formal design of dynamic reconfiguration protocol for cloud applications

Featuring edge caching and computing, fog radio access networks have been seen as promising architectures. However, the joint optimization of cache and radio resource can put heavy burden on the resource manager in the cloud. To overcome this issue, hierarchical radio and cache resource management is studied in this paper. The core idea is to fully utilize the resource management capabilities of fog access points (FAPs) to divide time-domain resource on a small timescale by participating a coalitional game while make the resource manager allocate cache resource on a large timescale to maximize a long-term utility. Based on defined FAP preference, a low-complexity and distributed coalition formation algorithm is first developed under per-FAP fronthaul capacity constraints. Then, facing the challenges incurred by no closed form, discrete variables and the curse of dimensionality, multi-agent reinforcement learning (MARL) based caching is proposed for the resource manager. In the proposal, multiple agents are created, one for each content-FAP pair, who jointly learn a caching strategy via the interaction with a history network environment. By simulation, the effectiveness of the MARL based caching is demonstrated.