Level Virtual Machine Research Articles

An empirical study on divergence of differently-sourced LLVM IRs

In solving binary code similarity detection, many approaches choose to operate on certain unified intermediate representations (IRs), such as Low Level Virtual Machine (LLVM) IR, to overcome the cross-architecture analysis challenge induced by the significant morphological and syntactic gaps across the diverse instruction set architectures (ISAs). However, the LLVM IRs of the same program can be affected by diverse factors, such as the acquisition source, i.e., compiled from source code or disassembled and lifted from binary code. While the impact of compilation settings on binary code has been explored, the specific differences between LLVM IRs from varied sources remain underexamined. To this end, we pioneer an in-depth empirical study to assess the discrepancies in LLVM IRs derived from different sources. Correspondingly, an extensive dataset containing nearly 98 million LLVM IR instructions distributed in 808,431 functions is curated with respect to these potential IR-influential factors. On this basis, three types of code metrics detailing the syntactic, structural, and semantic aspects of the IR samples are devised and leveraged to assess the divergence of the IRs across different origins. The findings offer insights into how and to what extent the various factors affect the IRs, providing valuable guidance for assembling a training corpus aimed at developing robust LLVM IR-oriented pre-training models, as well as facilitating relevant program analysis studies that operate on the LLVM IRs.

GAHLS: an optimized graph analytics based high level synthesis framework

The urgent need for low latency, high-compute and low power on-board intelligence in autonomous systems, cyber-physical systems, robotics, edge computing, evolvable computing, and complex data science calls for determining the optimal amount and type of specialized hardware together with reconfigurability capabilities. With these goals in mind, we propose a novel comprehensive graph analytics based high level synthesis (GAHLS) framework that efficiently analyzes complex high level programs through a combined compiler-based approach and graph theoretic optimization and synthesizes them into message passing domain-specific accelerators. This GAHLS framework first constructs a compiler-assisted dependency graph (CaDG) from low level virtual machine (LLVM) intermediate representation (IR) of high level programs and converts it into a hardware friendly description representation. Next, the GAHLS framework performs a memory design space exploration while account for the identified computational properties from the CaDG and optimizing the system performance for higher bandwidth. The GAHLS framework also performs a robust optimization to identify the CaDG subgraphs with similar computational structures and aggregate them into intelligent processing clusters in order to optimize the usage of underlying hardware resources. Finally, the GAHLS framework synthesizes this compressed specialized CaDG into processing elements while optimizing the system performance and area metrics. Evaluations of the GAHLS framework on several real-life applications (e.g., deep learning, brain machine interfaces) demonstrate that it provides 14.27× performance improvements compared to state-of-the-art approaches such as LegUp 6.2.

A Novel Data Management Scheme in Cloud for Micromachines

In cyber-physical systems (CPS), micromachines are typically deployed across a wide range of applications, including smart industry, smart healthcare, and smart cities. Providing on-premises resources for the storage and processing of huge data collected by such CPS applications is crucial. The cloud provides scalable storage and computation resources, typically through a cluster of virtual machines (VMs) with big data tools such as Hadoop MapReduce. In such a distributed environment, job latency and makespan are highly affected by excessive non-local executions due to various heterogeneities (hardware, VM, performance, and workload level). Existing approaches handle one or more of these heterogeneities; however, they do not account for the varying performance of storage disks. In this paper, we propose a prediction-based method for placing data blocks in virtual clusters to minimize the number of non-local executions. This is accomplished by applying a linear regression algorithm to determine the performance of disk storage on each physical machine hosting a virtual cluster. This allows us to place data blocks and execute map tasks where the data blocks are located. Furthermore, map tasks are scheduled based on VM performance to reduce job latency and makespan. We simulated our ideas and compared them with the existing schedulers in the Hadoop framework. The results show that the proposed method improves MapReduce performance in terms of job latency and makespan by minimizing non-local executions compared to other methods taken for evaluation.

Service separation assisted DDoS attack mitigation in cloud targets

DDoS (Distributed denial of service) attacks target on the availability of the victim services. DDoS attacks, being resource intensive attacks, create a heavy resource contention. In the state of the art, we found that resource isolation for legitimate users assists in maintaining service availability even in the presence of DDoS attacks. Resource isolation is an important method to minimize performance interference originated by the resource contention. In this paper, we propose resource isolation through service separation at two levels: physical machine level and virtual machine level. We compare four methods of resource segregation for legitimate users at physical machine level, virtual machine level, container level, and page level. The results of our experiments suggest that adopting different separation levels improve in the response time for legitimate user request even in the presence of DDoS attacks. By using different separation levels, we were able to see a considerable increase in the service availability time of target machines. In addition, there are notable improvements in number of requests failures and response time of target and co-located services.

Performance Analysis and Comparison of the Micro Virtual Machines Provided by the Top Cloud Vendors.

Public Clouds are the backbone of the modern IT Industry. The modern tech start-ups to are heavily dependent on the Public Clouds as save the capital expenditure. Public Clouds offer a wide range of Virtual Infrastructure beginning from single core 0.25 GB RAM virtual machines to 128 cored 3904 GB RAM Virtual Servers aided with GPUs. Apart from costing, The performance of the virtual machines can impact on the services offered. The paper tries to analyse the performance of the Light virtual machines provided by the different cloud vendors and maps their ability to handle certain set of tasks. The prime objective is to compare the entry level virtual machines that are mostly used across the industry for the multiple purposes and find which is the efficient and productive. The systems under Test are the 1 GB RAM machines provided by the most popular Cloud service providers.

Container Profiler: Profiling resource utilization of containerized big data pipelines.

This article presents the Container Profiler, a software tool that measures and records the resource usage of any containerized task. Our tool profiles the CPU, memory, disk, and network utilization of containerized tasks collecting over 60 Linux operating system metrics at the virtual machine, container, and process levels. The Container Profiler supports performing time-series profiling at a configurable sampling interval to enable continuous monitoring of the resources consumed by containerized tasks and pipelines. To investigate the utility of the Container Profiler, we profile the resource utilization requirements of a multistage bioinformatics analytical pipeline (RNA sequencing using unique molecular identifiers). We examine profiling metrics to assess patterns of CPU, disk, and network resource utilization across the different stages of the pipeline. We also quantify the profiling overhead of our Container Profiler tool to assess the impact of profiling a running pipeline with different levels of profiling granularity, verifying that impacts are negligible. The Container Profiler provides a useful tool that can be used to continuously monitor the resource consumption of long and complex containerized applications that run locally or on the cloud. This can help identify bottlenecks where more resources are needed to improve performance.

Machine learning based fast and accurate High Level Synthesis design space exploration: From graph to synthesis

In this paper, we present a machine learning based High-Level Synthesis (HLS) design space explorer (DSE) that significantly reduces the exploration runtime while leading to very accurate results. In order to do so, we leverage the power of low level virtual machine (LLVM) to generate the features used in the machine learning (ML) model. The proposed design space explorer uses a modified version of simulated annealing (SA) algorithm, where initially the search space is sampled to generate the predictive model. In this work we used gradient boost regression algorithm as our preferred ML model and achieve comparable results as a full DSE that performs logic synthesis for each newly generated design.

Intelligent Red Deer Algorithm based Energy Aware Load Balancing Scheme for Data Fusion in Cloud Environment

A cloud computing (CC) method was effectual if its sources were used in optimal way and an effectual consumption is attained by using and preserving proper management of cloud sources. Resource management can be attained through adoption of powerful source scheduling, allotment, and robust source scalability methods. The balancing of load in cloud is performed at VM level or physical machine level. A task use sources of VM and whenever a bunch of tasks reaches VM, the sources will be exhausted means no source is now existing for handling the extra task requests. This article develops an Intelligent Red Deer Algorithm based Energy Aware Load Balancing Scheme for data fusion in Cloud Environment, called IRDA-EALBS model. The presented IRDA-EALBS model majorly concentrates on the balancing of load among the virtual machines (VMs) in the cloud environment. The IRDA-EALBS model is mainly stimulated from the nature of red deers during a breading period. In addition, the IRDA-EALBS model derived an objective function to minimize energy consumption and maximize makespan. To demonstrate the enhanced performance of the IRDA-EALBS model, a wide range of experimental analyses is carried out. The simulation results highlighted the enhanced outcomes of the IRDA-EALBS model over other load balancers in the cloud environment.

Optimizing MapReduce Task Scheduling on Virtualized Heterogeneous Environments Using Ant Colony Optimization

Consuming Hadoop MapReduce via virtual infrastructure as a service is becoming common practice as cloud service providers (CSP) offers relevant applications and scalable resources. One of the predominant requirements of cloud users is to improve resource utilization in the virtual cluster during the service period. However, it may not be possible when MapReduce workloads and virtual machines (VM) are highly heterogeneous. Therefore, in this paper, we addressed these heterogeneities and proposed an efficient MapReduce scheduler to improve resource utilization by placing the right combination of the map and reduce tasks in each VM in the virtual cluster. To achieve this, we transformed the MapReduce task scheduling problem into a 2-Dimensional (2D) bin packing model and obtained an optimal schedule using the ant colony optimization (ACO) algorithm. As an added advantage, our proposed ACO based bin packing (ACO-BP) scheduler minimized the makespan for a batch of jobs. To showcase the performance improvement, we compared our proposed scheduler with three existing schedulers that work well in a heterogeneous environment. As expected, results show that ACO-BP significantly outperformed the existing schedulers while dealing with workload and VM level heterogeneities.

VM placement and cache based side channel attack at computational level in cloud computing : A Survey

There are different issues that need to focus in respect of security and privacy. One of the most security issue at computational level implementation of secure virtualization is cloud computing. In this, we explain virtualization security and their existing defense system at computational level. In computational level need to increase security of cloud and categorized the virtual machine level security challenges that are faced by cloud users at computational level in cloud computing. Day by day cloud attribute are introducing new challenges in cloud environment. In this paper, we focus on cloud security at computational level. In the day scenario to secure a cloud data on shared environment is the important task for IT industries. In virtual environment need to focus on secure VM placement on the host machine that is running on cloud. This paper demonstrates lack of VM placement and side channel attack (SCA). Analysis of different type of scheme is done and comparative study of game theory based, graph theory based and bin packing VM placement scheme. Also we discussed virtualization in cloud computing and their security aspects and comparative study of different type side channel attack (SCA).

Automated Software Vulnerability Detection Based on Hybrid Neural Network

Vulnerabilities threaten the security of information systems. It is crucial to detect and patch vulnerabilities before attacks happen. However, existing vulnerability detection methods suffer from long-term dependency, out of vocabulary, bias towards global features or local features, and coarse detection granularity. This paper proposes an automatic vulnerability detection framework in source code based on a hybrid neural network. First, the inputs are transformed into an intermediate representation with explicit structure information using lower level virtual machine intermediate representation (LLVM IR) and backward program slicing. After the transformation, the size of samples and the size of vocabulary are significantly reduced. A hybrid neural network model is then applied to extract high-level features of vulnerability, which learns features both from convolutional neural networks (CNNs) and recurrent neural networks (RNNs). The former is applied to learn local vulnerability features, such as buffer size. Furthermore, the latter is utilized to learn global features, such as data dependency. The extracted features are made up of concatenated outputs of CNN and RNN. Experiments are performed to validate our vulnerability detection method. The results show that our proposed method achieves excellent results with F1-scores of 98.6% and accuracy of 99.0% on the SARD dataset. It outperforms state-of-the-art methods.

A request aware module using CS-IDR to reduce VM level collateral damages caused by DDoS attack in cloud environment

Distributed Denial of Service (DDoS) plays a significant role in threatening the cloud-based services. DDoS is a kind of attack which targets the CPU, bandwidth and other resources and makes them unavailable to benign users. The DDoS attack has an enormous impact on multi-tenant cloud network than the traditional network due to the cloud features like virtualization, load balancing, resource scaling and migrations. These features spread attack effects in the whole cloud network, which introduces the collateral damages to the non-target stakeholders. Some of these stakeholders are co-hosted virtual machines (VMs), host physical server, co-hosted physical server, cloud service providers and users, etc. Therefore, there is a need for a method that can reduce such collateral damages. In this work, we focus on reducing VM level collateral damages caused to the co-hosted VMs residing with the victim VM on the same host. The proposed architecture consists of: (i) a request awareness based module to reduce VM level collateral damages, (ii) to obtain the request awareness, a novel Cuckoo Search based IDentification of Request (CS-IDR) method using bivariate flight is also proposed. The CS-IDR method helps in taking the request-aware decision, which eventually reduces VM level collateral damages. The result also shows that the proposed method minimizes the CPU usage, RAM usage, power consumption, overall load, and incurred cost caused due to DDoS attack on non-target co-hosted VM, and hence reduces such collateral damages.

Predictive Scaling for Elastic Compute Resources on Public Cloud Utilizing Deep Learning based Long Short-term Memory

The cloud resource usage has been increased exponentially because of adaptation of digitalization in government and corporate organization. This might increase the usage of cloud compute instances, resulting in massive consumption of energy from High performance Public Cloud Data Center servers. In cloud, there are some web applications which may experience diverse workloads at different timestamps that are essential for workload efficiency as well as feasibility of all extent. In cloud application, one of the major features is scalability in which most Cloud Service Providers (CSP) offer Infrastructure as a Service (IaaS) and have implemented auto-scaling on the Virtual Machine (VM) levels. Auto-scaling is a cloud computing feature which has the ability in scaling the resources based on demand and it assists in providing better results for other features like high availability, fault tolerance, energy efficiency, cost management, etc. In the existing approach, the reactive scaling with fixed or smart static threshold do not fulfill the requirement of application to run without hurdles during peak workloads, however this paper focuses on increasing the green tracing over cloud computing through proposed approach using predictive auto-scaling technique for reducing over-provisioning or under-provisioning of instances with history of traces. On the other hand, it offers right sized instances that fit the application to execute in satisfying the users through on-demand with elasticity. This can be done using Deep Learning based Time-Series LSTM Networks, wherein the virtual CPU core instances can be accurately scaled using cool visualization insights after the model has been trained. Moreover, the LSTM accuracy result of prediction is also compared with Gated Recurrent Unit (GRU) to bring business intelligence through analytics with reduced energy, cost and environmental sustainability.

REDUCER: Elimination of Repetitive Codes for Accelerated Iterative Compilation

Low Level Virtual Machine (LLVM) is a widely adopted open source compiler providing numerous optimization opportunities. The discovery of the best optimization sequence in this large space is done via iterative compilation, which incurs substantial overheads, especially for big data applications operating on high volume and variety datasets. The large search space is mostly comprised of identical codes generated via different optimizations. However, no mechanism is implemented inside the LLVM compiler to suppress the redundant testings. In this regard, this paper proposes REDUCER for eliminating the identical code executions by performing Intermediate Representation (IR) level comparisons. REDUCER has been tested using the well-accepted MiCOMP technique in LLVM 3.8 and 9.0 compiler, with embedded (cBench) and big data workloads. In comparison to MiCOMP 19.5 k experiments, REDUCER lowers the experiment count up to 327, i.e. 98 %, and on average to 4 375, i.e. 77 %, for cBench (LLVM-3.8). Similarly, for LLVM-9.0 the reductions are up to 1 931, i.e. 90 %, and on average 5 863, i.e. 69.9 %. Due to the significant experiment reduction, for embedded workloads, the iterative compilation is up to 58.6× and on average 4.1× faster with REDUCER (LLVM-3.8) than MiCOMP, whereas, with REDUCER (LLVM-9.0) the compilation is up to 8.5× and on average 2.9× faster. Moreover, REDUCER is found to be scalable and efficient for big data workloads where the iterative compilation is reduced to few days, as code is compared one time only for a single application tested on multiple datasets.

Resource provisioning in scalable cloud using bio-inspired artificial neural network model

Resource assignment is one of the emerging research area in the cloud scenario. Cloud computing provides a shared pool of resources in a distributed environment. It supports the features of utility-based computing. Efficient task provisioning on virtual machines is the major concern in an extensible cloud computing environment. The task provisioning minimizes the performance metrics total completion time (ms), average start time, average finish time, average execution time, scheduling time, and simulation time respectively. The scheduling is an important problem which becomes more complicated when various parameters consider. The key issue in virtual machine level scheduling is execution time overhead and scalability in a real-time scenario. Our objective is to make an optimal schedule of tasks on a virtual machine inside the datacenter using neural-bio inspired GA-ANN technique. This work presents a scheduler based on a genetic approach and an artificial neural network. The presented approach performs optimal scheduling of tasks on an appropriate virtual machine. The reliability of the system improves by reducing the number of tasks failed. The presented work uses a genetic algorithm to generated huge data sets and trains the neural model using the data set generated by using a genetic approach. The accuracy of the model is improved using back propagation with 98% accuracy. The set of experiments are performed using a scalable cloud computing environment. The presented bio-inspired technique is compared against nature-inspired, bio-inspired cost-aware BB–BC, GA-Cost, and GA-Exe based efficient task scheduling techniques. The results are obtained using real workload logs and synthetic data sets. Results indicate that the proposed GA-ANN bio-inspired predictive approach outperforms the considered nature-inspired scheduling approaches. The proposed algorithm is compared using various performance metrics total completion time, average start time, average finish time, and the fault rate, execution time, and scheduling time respectively. The proposed model reduces the fault rate by 82.63%, successfully completed tasks count improves by 26.81% and execution time improves by 10.66% and scheduling time improves by 69.94%. The scheduling time improves by 85.76% with an increasing number of iterations and constant numbers of tasks. Hence the presented GA-ANN scheduling technique outperformed the GA cost, GA EXE, and BB–BC COST scheduling approaches.

An adaptive workload-aware power consumption measuring method for servers in cloud data centers

As cloud computing technologies and applications develop rapidly in recent years, the quantity and size of cloud datacenters have been ever-increasing, making the overconsumption of energy in datacenters become a widespread concern. To reduce the energy cost by servers, we must first build an accurate power model to achieve flexible, device-free power consumption measuring. However, most of the previous work related to server power modeling solely apply to the server and virtual machine levels, and the existing power models fail to take into account the heterogeneity in workload. Therefore, we first propose separate power consumption models based on the distinction of workload types including CPU-intensive, I/O-intensive, memory-intensive, and mixed workload. Then, we present an adaptive workload-aware power consumption measuring method (WSPM) for cloud servers. Our method proactively selects an appropriate power model for the upcoming workload through workload clustering, forecasting and classification, which are implemented using K-means, ARIMA, and threshold-based methods, respectively. We conducted several experiments to evaluate the performance of the key components of our method. The result shows: (1) the accuracy of our future workload forecasting on real traces of requests to our servers, (2) the accuracy of the power consumption measured by WSPM, and (3) the effectiveness of our workload-aware method in reducing real-time power estimation lag. Overall, the proposed method simplifies power modeling under diverse workloads without losing accuracy, making it a general and highly available solution for cloud data centers.

Priority Based Mechanism for Resource Sharing in Cloud

Cloud computing technologies are getting matured day by day. Revolutions in underlying software, virtualization and hardware technologies related to storage, processing and computing technologies has helped cloud computing service providers to win trust of concerned stake holders. However, the exponentially increasing demand of cloud based resources has made task of resource management and utilization more and more challenging. A novel load balancing technique in cloud computing environment is presented in this paper. The virtual machines are implemented on an open source cloud computing platform on open source operating system. A virtual machines’ priorities based load balancing approach presented here indicates improvement in overall waiting time for load balancing. The mechanism prioritizes load balancing on same priority level virtual machines or lower priority level virtual machines.

IEEE Access Special Section Editorial: Emerging Trends, Issues, and Challanges in Energy-Efficient Cloud Computing

Cloud computing is one of the most successful business models for providing a simple pay-as-you-go, and therefore is gaining much popularity in the industry. Customers and enterprises can maintain or scale-up a business easily while cutting down on their budget. However, energy consumption is one of the biggest problems in current cloud computing. It is both essential and urgent for governmental and industrial institutions to address this, to achieve rapid growth. The development of energy-efficient cloud computing has to be taken into consideration, which relies on the development of several key technologies: More energy-efficient mediums can be used in cloud computing at the platform level; energy-efficient scheduling algorithms, memory systems, storage systems, resource management policies, etc., can be adopted at the hypervisor level; energy-efficient scheduling, communications, and applications can be applied at the virtual machine level; and energy-efficient mobile cloud computing will be developed, involving green networking and wireless communications, cloud-based mobile applications, and limited resources management.

Improving MapReduce scheduler for heterogeneous workloads in a heterogeneous environment

SummaryBig data is largely influencing business entities and research sectors to be more data‐driven. Hadoop MapReduce is one of the cost‐effective ways to process large scale datasets and offered as a service over the Internet. Even though cloud service providers promise an infinite amount of resources available on‐demand, it is inevitable that some of the hired virtual resources for MapReduce are left unutilized and makespan is limited due to various heterogeneities that exist while offering MapReduce as a service. As MapReduce v2 allows users to define the size of containers for the map and reduce tasks, jobs in a batch become heterogeneous and behave differently. Also, the different capacity of virtual machines in the MapReduce virtual cluster accommodate a varying number of map/reduce tasks. These factors highly affect resource utilization in the virtual cluster and the makespan for a batch of MapReduce jobs. Default MapReduce job schedulers do not consider these heterogeneities that exist in a cloud environment. Moreover, virtual machines in MapReduce virtual cluster process an equal number of blocks regardless of their capacity, which affects the makespan. Therefore, we devised a heuristic‐based MapReduce job scheduler that exploits virtual machine and MapReduce workload level heterogeneities to improve resource utilization and makespan. We proposed two methods to achieve this: (i) roulette wheel scheme based data block placement in heterogeneous virtual machines, and (ii) a constrained 2‐dimensional bin packing to place heterogeneous map/reduce tasks. We compared heuristic‐based MapReduce job scheduler against the classical fair scheduler in MapReduce v2. Experimental results showed that our proposed scheduler improved makespan and resource utilization by 45.6% and 47.9% over classical fair scheduler.

SIMD programming using Intel vector extensions

Single instruction multiple data (SIMD) extensions are one of the most significant capabilities of recent General Purpose Processors (GPPs) which improves the performance of applications with less hardware modification. Each GPP vendor such as HP, Sun, Intel, and AMD has its particular Instruction Set Architecture (ISA) and SIMD micro-architecture with different perspectives. Intel expanded SIMD technologies from hardware and software point of view. It has introduced SIMD technologies such as MultiMedia eXtensions (MMX), Streaming SIMD Extensions (SSE), Advanced Vector eXtensions (AVX), Fused Multiply Add (FMA) and AVX-512 sets. During micro-processors developments path, register width has been extended from 64 bits to 512 bits and number of vector registers has been increased from 8 to 32. Wider registers provide more parallelism ways and more registers reduce extra data movement to the cache memory. In order to gain the advantages of SIMD extensions, many programming approaches have been developed. Compiler Automatic Vectorization (CAV) as an implicit vectorization approach, provides simple and easy SIMDization tools. While, performance improvement of CAV is not always granted, most compilers auto-vectorize simple loops. On the other hand, for explicit vectorization, Intrinsic Programming Model (IPM) provides low-level access to vector registers for SIMDizing. However, programming with IPM requires great amount of expertise especially in low-level architecture feature, thus, choosing the suitable instructions and vectorization methodology for mapping to a certain algorithm is important. Moreover, portability, compatibility, scalability and compiler optimization might limit the advantage of IPM. Our goal in this paper is as follows. First, we provide a review of SIMD technology in general and Intel’s SIMD extensions in particular. Second, some SIMD features of Intel SIMD technologies, MMX, SSEs, AVX, and FMA in terms of ISA, vector width, and SIMD programming tools are comparatively discussed. Third, in order to compare the performance of different auto-vectorizers and IPM approaches using Intel C++ compiler (ICC), GNU Compiler Collection (GCC) and Low Level Virtual Machine (LLVM), we map and implement some representative multimedia kernels on AVX and AVX2 extensions. Finally, our experimental results show that although the performance improvement using IPM approach is higher than CAVs, programmer needs more programming efforts and knows different mapping strategists. Therefore, extending auto-vectorizers abilities to generate more efficient vectorized codes is an important issue in different compilers.