Hardware-based Solutions Research Articles

Camera calibration based on lightweight fan-shaped target detection and fitness-distance-balance chaotic marine predators algorithm

Camera calibration represents a critical stage in visual measurement, directly impacting the measurement precision. Addressing the limitations of checkerboard and circular targets, resulting in low calibration accuracy and eccentricity errors, this study introduces a camera calibration method that relies on lightweight fan-shaped target detection and Fitness-Distance-Balance Chaotic Marine Predators Algorithm(FDBCMPA). Starting with the improvement of camera calibration targets by refining and designing fan-shaped targets, this paper aims to tackle eccentricity errors through hardware-based solutions. In the process of extracting target feature points, a lightweight object detection algorithm is devised to capture individual fan-shaped pattern positions from images, transitioning from global multi-shape feature detection to localized single-pattern feature extraction. This approach mitigates the distortions inherent in traditional checkerboard targets during global linear feature imaging and partially alleviates interference from background environments and similar features. Following the initial camera calibration, the FDBCMPA is employed to further optimize the calibration outcomes, thereby boosting the calibration method's performance. Experimental findings highlight that the enhanced object detection method in this study attains detection accuracy comparable to mainstream methods, with distinct advantages in detection speed, parameters, and FLOPs, rendering it more suitable for use in camera calibration. Additionally, the FDBCMPA demonstrates improvements in convergence results compared to MPA, PSO, TACPSO, Chimp, and other methods. The final camera calibration results demonstrate an impressive reprojection error of 0.0738 pixels through the utilization of the proposed camera calibration method introduced in this paper, and substantiate the practicality and effectiveness of the calibration method proposed in this study.

Secure Cloud Architecture

Cloud computing refers to the availability of online resources and services. Global data centers serve as the delivery hubs for cloud services. Cloud computing helps its users by giving them access to virtual resources over the internet. Pay-Per Use-On-Demand mode is provided by cloud computing and e-trade. It may be easily accessible and shared via IT resources with the help of the internet. The primary obstacle in cloud computing is to protection and privacy issues arising from its multi-tenancy structure, as well as the outsourcing of critical applications, infrastructure, and sensitive data. Businesses are utilizing cloud services at a rapid pace; therefore, it is necessary to implement measures that ensure businesses have security and can select a provider that suits them. An advancement in current technology that fulfills the long-held dream of computing as a utility is cloud computing. The majority of private and public sector businesses have been deceived by the emergence of this innovative technology in the IT industry. The two main security-based approaches now in use for cloud-based platforms are holomorphic encryption and single tamper-proof hardware. Scalability is a problem for hardware-based solutions, but holomorphic encryptions are merely an idea. Furthermore, because of the different nature of its services and deployment strategy, cloud-based platforms cannot directly incorporate standard defense in-depth security mechanisms. To secure the cloud-based platform, however, the same idea of a multi-layered security mechanism can be suggested

Model-based deep learning for ultrasound beamforming

Recently, the ultrasound signal processing pipeline has shifted from hardware-based solutions to the digital domain, enabling more intricate reconstruction techniques. We highlight several beamforming methods from a statistical inference perspective, connecting deep learning techniques to established signal processing models grounded in fundamental principles. The ultrasound acquisition process can be modeled as a linear combination of scatterers x (i.e. tissue reflectivity) which amounts to the received channel data y, captured by a forward model y = Hx + n, with the measurement matrix H. Solving this system relies heavily on priors to yield a unique and anatomically feasible solution. A naive linear estimator for x is H T y, also known as the delay-and-sum beamformer, which crudely disregards all off-axis scattering as zero-mean white noise. Based on the structured-noise assumption for off-axis scattering in the minimum variance beamformer, ABLE implements fast statistical inference of the optimal linear beamformer through deep learning. In addition, we can infuse prior knowledge on the distribution of x resulting in the neural maximum-a-posteriori beamformer. All aforementioned methods assume no dependency between individual pixels. Ultimately, the use of generative models allows us to extend the methods with more informative spatial priors.

Driver-Activated VIDeoscopic Surgery: An Innovation Journey from Idea to Industry-Ready.

Introduction: Innovation is not a straightforward path. While surgeons are intimately familiar with clinical problems and often devise clever solutions to address them, the journey from idea to a marketable product is opaque. We describe our experience developing a novel video navigation system to help streamline collaboration and enhance surgeon control of their video image in minimally invasive surgery. Materials and Methods: Our idea began with recognizing the primary clinical challenge: "one bad image for all" in laparoscopic surgery, when the least experienced member of the surgical team is often expected to hold the camera. Results: Through multiple iterations and pivots, including hardware-based solutions like head-mounted displays and individualized monitors, we arrived at a hardware-agnostic software algorithm to process laparoscopic video for real-time image navigation. As we explain why, how, and when to pivot, we provide brief overviews of protecting intellectual property and financing innovation. Finally, collaboration with professional societies, such as the International Pediatric Endosurgery Group, provides fertile testing grounds for new ideas. Conclusion: Our experience may help future surgeon-innovators go from their ideas to industry-ready.

Optimized real-time parking management framework using deep learning

Today’s modern world has seen massive population growth in the past few decades, causing the automobile industry to expand rapidly. The exponential increase has disrupted the seamless flow of traffic globally and has fashioned significant problems such as traffic congestion and its mismanagement. To tackle these issues, some hardware and data-driven-based solutions have been previously proposed. The hardware-based solutions utilize different types of sensors installed at relevant sites to monitor the status of parking slots which makes them less scalable, complex, costly to install and maintain. In contrast, the data-driven solutions utilize the already available infrastructure of surveillance cameras installed at parking lots, thus overcoming the limitations of sensor-based solutions. However, currently only classification-based approaches are prevalent and adopted to monitor the status of the parking slots which makes the systems less generalized for scalability with slower performance. This paper proposes an intelligent parking management system which employs deep learning to alleviate the limitations in the data driven solutions by leveraging the high performance and fast inference capability of YOLO v5 for vehicles detection instead of parking slot classification. The model was evaluated using the PKLot dataset which is a benchmark for identifying the status of the parking lot with state of the art performance achieved having an accuracy of 99.5%. To augment the performance of the algorithm in real-time, a pretrained model of YOLO v5 on MS COCO dataset was employed to detect and assign vacant parking slots and generate vehicle statistics. The performance of the proposed system was evaluated on our custom dataset with an accuracy of 96.8% achieved and a nearly real-time performance of 45 Fps which makes it more efficient, scalable, and generalized.

A parallel decision-making design for highly speedy packet classification

Packet classification is crucial in computer networks to increase network security because of developments in high-speed data communication. To support different network services including Quality of Service (QoS), security, and resource reservation, network packet classification is a crucial network kernel function. It became extremely challenging to classify arriving packets using the traditional packet classification algorithms at a decent pace due to the rapidly increasing size of rulesets and rule fields in current networks. In addition to hardware-based solutions, numerous contemporary software-based classification techniques have been put out to speed up packet classification. In general, it's critical to manage low latency, fast throughput, and higher energy efficiency with minimal memory needs while designing a packet classification method. In this paper, proposed the architecture makes use of the parallelism provided by current hardware technologies to classify numerous packets at once to improve throughput, reduce energy consumption, and significantly decrease the latency. The suggested high-speed parallel classification procedure uses two DPRAMs (DPRAM1 and DPRAM2) to support multiple read of the data. DPRAMs(Dual Port Random Access Memories) are preferred for stage memory because of their low latency performance, which enables quick rate packet lookup. The packet scheduling module receives the data from the DPRAM. The four successive packet splitter modules receive the packets in parallel. As a result, the packet classification throughput increased. To optimize throughput, and latency while maintaining low energy consumption, facilitating parallel packet classification architecture on FPGA is discussed in this work. Several parameters, including throughput, latency, energy efficiency, and memory requirements, will be evaluated and compared to traditional methods to assess performance. The classifier offers less latency of 51 ns and the lowest energy efficiency of 5.2 nJ, and its throughput may approach 802 million packets per second at the rate of 420 MHz clock frequency.

Enabling Fast and Memory-Efficient Acceleration for Pattern Matching Workloads: The Lightweight Automata Processing Engine

Growing pattern matching applications are employing finite automata as their basic processing model. These applications match tens to thousands of patterns on a large amount of data, which brings a great challenge to conventional processors. Therefore hardware-based solutions have emerged frequently and achieved high throuphput automata processing. However, existing methods are generally difficult to achieve both processing speed and storage efficiency, and are often too heavy to be integrated into a small chip and have to rely on off-chip DRAMs or other high capacity memories even on some simple data sets, leading to the potential area and power consumption issues. In this paper, we focus on building a more lightweight automata processing engine, hoping to store the whole automata model into on-chip memory and run effectively and independently. We propose LAP, a lightweight automata processing engine. Powered with a novel automata model (A-DFA) and efficient packing algorithms, extremely high storage efficiency compared with traditional DFA is achieved in LAP. Meanwhile, we identify the key parallelization factors in the A-DFA model and then propose a specialized microarchitecture with novel instructions to further accelerate the state transition process. As a result, LAP can obtain more effective trade-off between processing speed and storage efficiency. Evaluation results show that LAP achieves extremely high storage efficiency on simple data sets, exceeding IBM's RegX by 8×, and achieves significant improvements in processing speed ranging from 1.32× to 1.91× compared with previous lightweight hardware implementations. Moreover, LAP has good scalability in hardware architecture. It is easy to build an acceleration system with higher throughput by increasing the number of cores. We prototype a 16-core system into Xilinx ZC702 FPGA and a 64-core system into Xilinx ZCU102 FPGA respectively. The prototype system on ZC702 on average achieves 3.5 GB/s throughput on simple data sets, and the prototype system on ZCU102 can obtain higher throughput and compute density values on part of large datasets in ANMLZoo compared with modern in-memory NFA-based solutions.

A survey on face presentation attack detection mechanisms: hitherto and future perspectives.

The advances in human face recognition (FR) systems have recorded sublime success for automatic and secured authentication in diverse domains. Although the traditional methods have been overshadowed by face recognition counterpart during this progress, computer vision gains rapid traction, and the modern accomplishments address problems with real-world complexity. However, security threats in FR-based systems are a growing concern that offers a new-fangled track to the research community. In particular, recent past has witnessed ample instances of spoofing attacks where imposter breaches security of the system with an artifact of human face to circumvent the sensor module. Therefore, presentation attack detection (PAD) capabilities are instilled in the system for discriminating genuine and fake traits and anticipation of their impact on the overall behavior of the FR-based systems. To scrutinize exhaustively the current state-of-the-art efforts, provide insights, and identify potential research directions on face PAD mechanisms, this systematic study presents a review of face anti-spoofing techniques that use computational approaches. The study includes advancements in face PAD mechanisms ranging from traditional hardware-based solutions to up-to-date handcrafted features or deep learning-based approaches. We also present an analytical overview of face artifacts, performance protocols, and benchmark face anti-spoofing datasets. In addition, we perform analysis of the twelve recent state-of-the-art (SOTA) face PAD techniques on a common platform using identical dataset (i.e., REPLAY-ATTACK) and performance protocols (i.e., HTER and ACA). Our overall analysis investigates that despite prevalent face PAD mechanisms demonstrate potential performance, there exist some crucial issues that requisite a futuristic attention. Our analysis put forward a number of open issues such as; limited generalization to unknown attacks, inadequacy of face datasets for DL-models, training models with new fakes, efficient DL-enabled face PAD with smaller datasets, and limited discrimination of handcrafted features. Furthermore, the COVID-19 pandemic is an additional challenge to the existing face-based recognition systems, and hence to the PAD methods. Our motive is to present a complete reference of studies in this field and orient researchers to promising directions.

Evaluation of a two-dimensional Moire-free antiscatter grid for cone-beam computed tomography.

Scatter radiation is a traditional problem in cone-beam computed tomography (CBCT). Accordingly, numerous methods have been researched for scatter reduction in terms of software- or hardware-based solutions. The concept of a two-dimensional antiscatter grid (2D-ASG) has been shown to provide a solution to the scatter reduction in CBCT. However, the use of an ASG makes it challenging to use in clinical CBCT systems because it causes Moire artifacts of theimage. We have developed a Moire-free 2D-ASG that was designed to solve the Moire artifact and the scatter radiation problems. We provide the experimental results pertaining to the image quality measurements from the evaluation of the 2D-ASG compared with those obtained from a one-dimensional antiscatter grid (1D-ASG) and no antiscatter grid (No-ASG) to demonstrate the quantitative extent of the improvements. The 2D-ASG, fabricated based on a sawing process with a graphite body, was prepared to evaluate image quality improvements. Projection images for Pro-CT MK II phantom were acquired using the CBCT testbed of sample rotation type and reconstructed by Feldkamp-Davis-Kress (FDK) algorithm without any filters. We measured the contrast-to-noise ratio (CNR) and uniformity index for the cupping artifacts of the 2D-ASG, 1D-ASG, and No-ASG cases. 2D-ASG considerably reduced the cupping artifacts owing to the scatter reduction compared with the 1D-ASG and No-ASG cases. The cupping artifacts were reduced by 85% in the 2D-ASG compared with No-ASG, while the cupping artifacts were reduced by 63% in 1D-ASG compared with No-ASG. The 2D-ASG also yielded a CNR improvement. The average CNR improvements for eight insert materials were 47% in the 2D-ASG compared with No-ASG, while the CNR was improved by 36% in the 1D-ASG compared withNo-ASG. We demonstrated that the Moire-free 2D-ASG improved the image quality by removing scatter radiation in CBCT compared with 1D-ASG and No-ASG. We believe that the Moire-free 2D-ASG can become one of the effective ways to solve the scatter radiation problem in CBCT images because it provides usability and has the potential to have synergistic effects on other methods, such as bow-tie filter and scattercorrection.

A Survey on harnessing the Applications of Mobile Computing in Healthcare during the COVID-19 Pandemic: Challenges and Solutions

The COVID-19 pandemic ravaged almost every walk of life but it triggered many challenges for the healthcare system, globally. Different cutting-edge technologies such as Internet of things (IoT), machine learning, Virtual Reality (VR), Big data, Blockchain etc. have been adopted to cope with this menace. In this regard, various surveys have been conducted to highlight the importance of these technologies. However, among these technologies, the role of mobile computing is of paramount importance which is not found in the existing literature. Hence, this survey in mainly targeted to highlight the significant role of mobile computing in alleviating the impacts of COVID-19 in healthcare sector. The major applications of mobile computing such as software-based solutions, hardware-based solutions and wireless communication-based support for diagnosis, prevention, self-symptom reporting, contact tracing, social distancing, telemedicine and treatment related to coronavirus are discussed in detailed and comprehensive fashion. A state-of-the-art work is presented to identify the challenges along with possible solutions in adoption of mobile computing with respect to COVID-19 pandemic. Hopefully, this research will help the researchers, policymakers and healthcare professionals to understand the current research gaps and future research directions in this domain. To the best level of our knowledge, this is the first survey of its type to address the COVID-19 pandemic by exploring the holistic contribution of mobile computing technologies in healthcare area.

Product Development Process for complex hardware-based solutions: current trends

Innovation processes have evolved significantly. Starting with traditional Product Development Process in the 1990s, novel development approaches have been disseminated, especially approaches originating in the software and service sectors. However, understanding the influence of recent approaches to developing complex hardware-based solutions is still a research issue. This study aims to identify emerging approaches that have already demonstrated adequacy to support hardware design. Therefore, a systematic literature review has been conducted. The research screened 292 papers and analyzed in detail 25 articles published between 2018 and 2022. Our findings indicate increased adoption of creativity-oriented methods and extended prototyping in early phases.

Classification of random number generator applications in IoT: A comprehensive taxonomy

Randomness is an important issue for Internet of Things (IoT). The need to generate suitable random numbers for IoT devices with resource and size limitations has emerged due to the cryptographic protocols. Although random number generation approaches have been proposed considering IoT device constraints, commonly used software and hardware-based solutions have not been discussed in detail. The main contribution of this paper is the detailed examination of the problems encountered in random number generation in the IoT ecosystem and the proposed solution approaches. In this context, a classification has been proposed for hardware containing random number generator (RNG), which has different usage areas in IoT environments. Based on the presented classification, the characteristics of the devices in terms of resource constraints are examined. This classification serves as a guide for selecting suitable hardware in applications with or without random number needs. Also, basic RNGs and test suites are discussed. Some challenges are summarized by explaining the random numbers usage in the IoT environment. In addition, proposed random number generation scenarios for IoT devices are determined. Success rate analysis is carried out based on these techniques in terms of randomness tests. Software and hardware-based solution methods are detailed to meet the need for random numbers in real-world IoT applications. RNG algorithms actively used in IoT applications, basic working principles, usage areas, and hardware features are summarized. Finally, the problem that arises in generating random numbers in system-on-chip (SoC) systems, one of the proposed classification components, are summarized, and some precautions that can be taken are expressed.

Motion artifacts removal and evaluation techniques for functional near-infrared spectroscopy signals: A review.

With the emergence of an increasing number of functional near-infrared spectroscopy (fNIRS) devices, the significant deterioration in measurement caused by motion artifacts has become an essential research topic for fNIRS applications. However, a high requirement for mathematics and programming limits the number of related researches. Therefore, here we provide the first comprehensive review for motion artifact removal in fNIRS aiming to (i) summarize the latest achievements, (ii) present the significant solutions and evaluation metrics from the perspective of application and reproduction, and (iii) predict future topics in the field. The present review synthesizes information from fifty-one journal articles (screened according to three criteria). Three hardware-based solutions and nine algorithmic solutions are summarized, and their application requirements (compatible signal types, the availability for online applications, and limitations) and extensions are discussed. Five metrics for noise suppression and two metrics for signal distortion were synthesized to evaluate the motion artifact removal methods. Moreover, we highlight three deficiencies in the existing research: (i) The balance between the use of auxiliary hardware and that of an algorithmic solution is not clarified; (ii) few studies mention the filtering delay of the solutions, and (iii) the robustness and stability of the solution under extreme application conditions are not discussed.

Deep Learning Based Phishing Websites Detection

Phishing is a crime that involves the theft of confidential user information. Those targeted by phishing websites include individuals, small businesses, cloud storage providers, and government organisations and websites. The majority of phishing prevention techniques involve hardware-based solutions, although software-based options are preferred due of cost and operational considerations. There is no answer to the problem of zero-day phishing assaults from the present phishing detection approaches since there is no solution to the problem. The Phishing Attack Detector based on Web Crawler, a three-phase attack detection system, was designed to handle these issues and accurately detect phishing incidences using a recurrent neural network in order to resolve these issues. It covers the input aspects of web traffic, web content, and Uniform Resource Locator (URL) based on the classification of phishing and non-phishing pages, as well as the output features of phishing and non-phishing pages.

Trusted Platform Module-Based Privacy in the Public Cloud: Challenges and Future Perspective

Public cloud providers offer ready-to-use, easily scalable servers on demand for a variety of applications. Storing and processing private and sensitive data in the cloud brings multiple security issues and indeed these concerns currently prevent many users from utilizing cloud resources. Improving both security and trust for users is increasingly important for cloud providers. In this context, we first investigate the classes of security threats encountered by cloud applications. We then analyze various software- and hardware-based solutions to handle these security challenges and provide the user with a chain of trust. Our analysis shows that pure software-based solutions do not sufficiently mitigate the challenges of the cloud environment. Hardware-based solutions utilizing the Trusted Platform Module (TPM) alleviate the issues; however, it is challenging to implement in the public cloud environment. Finally, we introduce our TPM-SGX-based approach that utilizes software TPM and Software Guard Extension (SGX) to provide similar security as the hardware-based approach.

Polarization Aberrations in High-Numerical-Aperture Lens Systems and Their Effects on Vectorial-Information Sensing

The importance of polarization aberrations has been recognized and studied in numerous optical systems and related applications. It is known that polarization aberrations are particularly crucial in certain photogrammetry and microscopy techniques that are related to vectorial information—such as polarization imaging, stimulated emission depletion microscopy, and structured illumination microscopy. Hence, a reduction in polarization aberrations would be beneficial to different types of optical imaging/sensing techniques with enhanced vectorial information. In this work, we first analyzed the intrinsic polarization aberrations induced by a high-NA lens theoretically and experimentally. The aberrations of depolarization, diattenuation, and linear retardance were studied in detail using the Mueller matrix polar-decomposition method. Based on an analysis of the results, we proposed strategies to compensate the polarization aberrations induced by high-NA lenses for hardware-based solutions. The preliminary imaging results obtained using a Mueller matrix polarimeter equipped with multiple coated aspheric lenses for polarization-aberration reduction confirmed that the conclusions and strategies proposed in this study had the potential to provide more precise polarization information of the targets for applications spanning across classical optics, remote sensing, biomedical imaging, photogrammetry, and vectorial optical-information extraction.

Efficient pipelined flow classification for intelligent data processing in IoT

The packet classification is a fundamental process in provisioning security and quality of service for many intelligent network-embedded systems running in the Internet of Things (IoT). In recent years, researchers have tried to develop hardware-based solutions for the classification of Internet packets. Due to higher throughput and shorter delays, these solutions are considered as a major key to improving the quality of services. Most of these efforts have attempted to implement a software algorithm on the FPGA to reduce the processing time and enhance the throughput. The proposed architectures, however, cannot reach a compromise among power consumption, memory usage, and throughput rate. In view of this, the architecture proposed in this paper contains a pipeline-based micro-core that is used in network processors to classify packets. To this end, three architectures have been implemented using the proposed micro-core. The first architecture performs parallel classification based on header fields. The second one classifies packets in a serial manner. The last architecture is the pipeline-based classifier, which can increase performance by nine times. The proposed architectures have been implemented on an FPGA chip. The results are indicative of a reduction in memory usage as well as an increase in speedup and throughput. The architecture has a power consumption of is 1.294w, and its throughput with a frequency of 233 ​MHz exceeds 147 Gbps.

Vehicle Detection for Vision-Based Intelligent Transportation Systems Using Convolutional Neural Network Algorithm

Vehicle detection in Intelligent Transportation Systems (ITS) is a key factor ensuring road safety, as it is necessary for the monitoring of vehicle flow, illegal vehicle type detection, incident detection, and vehicle speed estimation. Despite the growing popularity in research, it remains a challenging problem that must be solved. Hardware-based solutions such as radars and LIDAR are been proposed but are too expensive to be maintained and produce little valuable information to human operators at traffic monitoring systems. Software based solutions using traditional algorithms such as Histogram of Gradients (HOG) and Gaussian Mixed Model (GMM) are computationally slow and not suitable for real-time traffic detection. Therefore, the paper will review and evaluate different vehicle detection methods. In addition, a method of utilizing Convolutional Neural Network (CNN) is used for the detection of vehicles from roadway camera outputs to apply video processing techniques and extract the desired information. Specifically, the paper utilized the YOLOv5s architecture coupled with k-means algorithm to perform anchor box optimization under different illumination levels. Results from the simulated and evaluated algorithm showed that the proposed model was able to achieve a mAP of 97.8 in the daytime dataset and 95.1 in the nighttime dataset.

A Review on Software-Based and Hardware-Based Authentication Mechanisms for the Internet of Drones

During the last few years, a wide variety of Internet of Drones (IoD) applications have emerged with numerous heterogeneous aerial and ground network elements interconnected and equipped with advanced sensors, computation resources, and communication units. The evolution of IoD networks presupposes the mitigation of several security and privacy threats. Thus, robust authentication protocols should be implemented in order to attain secure operation within the IoD. However, owing to the inherent features of the IoD and the limitations of Unmanned Aerial Vehicles (UAVs) in terms of energy, computational, and memory resources, designing efficient and lightweight authentication solutions is a non-trivial and complicated process. Recently, the development of authentication mechanisms for the IoD has received unprecedented attention. In this paper, up-to-date research studies on authentication mechanisms for IoD networks are presented. To this end, the adoption of conventional technologies and methods, such as the widely used hash functions, Public Key Infrastructure (PKI), and Elliptic-Curve Cryptography (ECC), is discussed along with emerging technologies, including Mobile Edge Computing (MEC), Machine Learning (ML), and Blockchain. Additionally, this paper provides a review of effective hardware-based solutions for the identification and authentication of network nodes within the IoD that are based on Trusted Platform Modules (TPMs), Hardware Security Modules (HSMs), and Physically Unclonable Functions (PUFs). Finally, future directions in these relevant research topics are given, stimulating further work.

A review paper on hypervisor and virtual machine security

A hypervisor is a computer programme or software that facilitates to create and run multiple virtual machines. It is also known as Virtual Machine Manager (VMM). Due to their popularity, it exploits the attack surface, because the Hypervisor code contains much vulnerability. Since the Hypervisor is a core element of any cloud computing service, it is always on the top priority of the attackers. There are many Software (open source) and Hardware-based Solutions are available in the market to monitor and control the hypervisor activities. This paper summarizes various types of attacks, vulnerabilities, security issues and challenges related to hypervisor and virtual machines.