Terms Of CPU Utilization Research Articles

Deep Reinforcement Learning for Workload Prediction in Federated Cloud Environments.

The Federated Cloud Computing (FCC) paradigm provides scalability advantages to Cloud Service Providers (CSP) in preserving their Service Level Agreement (SLA) as opposed to single Data Centers (DC). However, existing research has primarily focused on Virtual Machine (VM) placement, with less emphasis on energy efficiency and SLA adherence. In this paper, we propose a novel solution, Federated Cloud Workload Prediction with Deep Q-Learning (FEDQWP). Our solution addresses the complex VM placement problem, energy efficiency, and SLA preservation, making it comprehensive and beneficial for CSPs. By leveraging the capabilities of deep learning, our FEDQWP model extracts underlying patterns and optimizes resource allocation. Real-world workloads are extensively evaluated to demonstrate the efficacy of our approach compared to existing solutions. The results show that our DQL model outperforms other algorithms in terms of CPU utilization, migration time, finished tasks, energy consumption, and SLA violations. Specifically, our QLearning model achieves efficient CPU utilization with a median value of 29.02, completes migrations in an average of 0.31 units, finishes an average of 699 tasks, consumes the least energy with an average of 1.85 kWh, and exhibits the lowest number of SLA violations with an average of 0.03 violations proportionally. These quantitative results highlight the superiority of our proposed method in optimizing performance in FCC environments.

Research on Edge Cloud Load Balancing Strategy Based on Chaotic Hierarchical Gene Replication

Edge cloud is used to handle latency-sensitive services. However, due to the large number of concurrent requests for edge intensive tasks, the resource allocation strategy affects the stability of nodes. In addition to an adaptive resource allocation model based on chaotic hierarchical gene replication (CRPSO model), the concept of chaotic replication ratio is proposed. This study is divided into two parts. The first is to verify the algorithm verification of the simulation platform. By comparison, it is found that CRPSO reduces the CPU and bandwidth utilization by 43.7% and 62.7% on average, respectively, and the memory usage is also lower than other algorithms. Thereafter, we compared the CRPSO algorithm with the Kubernetes clustering algorithm. Experiments showed that the fitness of the CRPSO model is 33.7% higher than that of the comparison algorithm on average. The algorithm is superior to the cluster scheduling algorithm in terms of CPU utilization and memory utilization. Furthermore, the total variance of the two resources involved in this model improved significantly, reaching 69.8% on average. In addition, CRPSO also has great advantages in other aspects of CPU and memory. It is indicated that the model in this study is suitable for the scenario of edge large-scale requests.

Analysis of Cryptography Algorithms Implemented in Android Mobile Application

This paper evaluates the performances of numerous encryption algorithms on mobile devices running the Android operating system. The primary objective of our research was to measure and compare the relative performances of tested algorithm implementations (Data Encryption Standard (DES), 3DES, Advanced Encryption Standard (AES), ChaCha20, Blowfish, and Rivest Cipher 4 (RC4)) on the Android platform. The algorithms were compared in terms of CPU utilization by measuring the time required to encrypt and decrypt variable size text files. Besides evaluating the six common symmetric encryption ciphers, a comparison has been conducted for several Password-Based Encryption (PBE) algorithms. Diverse cipher transformations were evaluated for each algorithm by utilizing various feedback modes and padding schemes. Two smartphone devices were used for testing, with different versions of the Android operating system and hardware specifications. The summarized performance outcomes for various cipher transformations are presented to demonstrate the effectiveness of each algorithm.

Performance optimization of task intensive real‐time applications on multicore ECUs—A hybrid scheduler

In the current approach of automotive electronic system design, the multicore processors have prevailed to achieve high computing performance at low thermal dissipation. Multicore processors offer functional parallelism that helps in meeting the safety critical requirements of vehicles. The number of Electronic Control Units (ECUs) in high-end cars could be reduced by conglomerating more functions into a multicore ECU. AUTOSAR stack has been designed to support the applications developed for multicore ECUs. The real challenges lie in adapting new design methods while developing sophisticated applications with multicore constraints. It is imperative to utilize the most of multicore computational capability towards enhancing the overall performance of ECUs. In this context, the scheduling of the real-time multitasking software components by the operating system is one of the challenging issues to be addressed. Here, the state-of-the-art scheduling algorithm is reviewed and its merits and limitations are identified. A hybrid scheduler has been proposed, tested and compared with the state-of-the-art algorithm that offers better performance in terms of CPU utilization, average response time and deadline missing rate both in normal and high load conditions.

Horizontal Scaling for Containerized Application Using Hybrid Approach

Cloud computing is a paradigm to provide services to end-users through the Internet. The availability of services to end-users is dependent on various factors such as the availability of computing resources as well as the number of users to access those services. To manage the real-time fluctuating workload cloud providers use elasticity mechanisms. Elasticity is one of the important characteristics of cloud computing that dynamically allocates computing resources to manage the fluctuating workload. The failure of allocation/de-allocation of computing resources at the right moment leads to SLA violation, degradation of services performance, maximum power consumption, minimum throughput, and maximum response time. To address these challenges, we have proposed a hybrid approach to perform horizontal elasticity. The proposed approach uses both reactive and proactive approaches for provisioning/de-provisioning of computing resources. The simulation results of the proposed model show that performance of system has improved in terms of CPU utilization, response time, and throughput.

CONSENSUS ALGORITHMS BASED BLOCKCHAIN OF THINGS FOR DISTRIBUTED HEALTHCARE

Internet of Things (IoT) consists of smart Things with evolution of ubiquitous computing. Fog Computing (FC) processes and analyzes data of these sensors near to users. However, the ever-increasing in number of Things and the consequent explosion in data traffic have led to fail traditional solutions of centralized storage. Blockchain is a new technology developed as a shared ledger build around peer-to-peer network to produce unchangeable blocks that contain multiple data. These blocks are linked to previous ones in sequence called chain through hash functions. Participants in blockchain selects a leader through one of consensus algorithms who adds new blocks in the chain to prevent dishonest nodes from creating invalid blocks. In this paper, we propose IoT based blockchain architecture named blockchain of Things to store medical records in a distributed manner. The architecture is emulated on Fog server Linux-based using Node.js and Postman. Three consensus protocols, namely: Proof of Work (PoW), Proof of Stake (PoS), and Byzantine Fault Tolerance (PBFT) are emulated and communicated using WebSocket. The consensus algorithms are evaluated in terms of CPU utilization and memory usage. The results show that PoS compared to PoW and PBFT is more lightweight and beneficial to IoT application.

Dynamic Scalability Model for Containerized Cloud Services

Cloud computing has become an important research area in large-scale computing systems and is being employed by many organizations in government, businesses, and industry. Schemes and appropriate models for dynamic resources provisioning in the cloud environment have been extensively studied. To date, the research literature is lacking schemes and models that offer dynamic scalability in which Quality of Service (QoS) and high performance are provided to customers with the usage of the least number of cloud resources, especially for containerized services hosted on the cloud. With dynamic scalability, cloud computing can offer on-demand, timely, and dynamically adjustable computing resources to services hosted on the cloud. This paper presents a dynamic scaling model based on queueing theory to scale containers virtual resources and satisfy the customer Service Level Agreements (SLA) while guarding costs of scaling very low. The aim is to improve the virtual computing resources utilization and satisfy SLA constraints in terms of CPU utilization, system response time, system drop rate, system number of tasks, and system throughput. Simulation results are provided using Java Modelling Tools simulation tool, which shows that our proposed model can determine under any offered workload the needed containers instances to satisfy the required QoS parameters.

Towards a Secure Thermal-Energy Aware Routing Protocol in Wireless Body Area Network Based on Blockchain Technology.

The emergence of biomedical sensor devices, wireless communication, and innovation in other technologies for healthcare applications result in the evolution of a new area of research that is termed as Wireless Body Area Networks (WBANs). WBAN originates from Wireless Sensor Networks (WSNs), which are used for implementing many healthcare systems integrated with networks and wireless devices to ensure remote healthcare monitoring. WBAN is a network of wearable devices implanted in or on the human body. The main aim of WBAN is to collect the human vital signs/physiological data (like ECG, body temperature, EMG, glucose level, etc.) round-the-clock from patients that demand secure, optimal and efficient routing techniques. The efficient, secure, and reliable designing of routing protocol is a difficult task in WBAN due to its diverse characteristic and restraints, such as energy consumption and temperature-rise of implanted sensors. The two significant constraints, overheating of nodes and energy efficiency must be taken into account while designing a reliable blockchain-enabled WBAN routing protocol. The purpose of this study is to achieve stability and efficiency in the routing of WBAN through managing temperature and energy limitations. Moreover, the blockchain provides security, transparency, and lightweight solution for the interoperability of physiological data with other medical personnel in the healthcare ecosystem. In this research work, the blockchain-based Adaptive Thermal-/Energy-Aware Routing (ATEAR) protocol for WBAN is proposed. Temperature rise, energy consumption, and throughput are the evaluation metrics considered to analyze the performance of ATEAR for data transmission. In contrast, transaction throughput, latency, and resource utilization are used to investigate the outcome of the blockchain system. Hyperledger Caliper, a benchmarking tool, is used to evaluate the performance of the blockchain system in terms of CPU utilization, memory, and memory utilization. The results show that by preserving residual energy and avoiding overheated nodes as forwarders, high throughput is achieved with the ultimate increase of the network lifetime. Castalia, a simulation tool, is used to evaluate the performance of the proposed protocol, and its comparison is made with Multipath Ring Routing Protocol (MRRP), thermal-aware routing algorithm (TARA), and Shortest-Hop (SHR). Evaluation results illustrate that the proposed protocol performs significantly better in balancing of temperature (to avoid damaging heat effect on the body tissues) and energy consumption (to prevent the replacement of battery and to increase the embedded sensor node life) with efficient data transmission achieving a high throughput value.

An efficient resource provisioning approach for analyzing cloud workloads: a metaheuristic-based clustering approach

With the recent advancements in Internet-based computing models, the usage of cloud-based applications to facilitate daily activities is significantly increasing and is expected to grow further. Since the submitted workloads by users to use the cloud-based applications are different in terms of quality of service (QoS) metrics, it requires the analysis and identification of these heterogeneous cloud workloads to provide an efficient resource provisioning solution as one of the challenging issues to be addressed. In this study, we present an efficient resource provisioning solution using metaheuristic-based clustering mechanism to analyze cloud workloads. The proposed workload clustering approach used a combination of the genetic algorithm and fuzzy C-means technique to find similar clusters according to the user’s QoS requirements. Then, we used a gray wolf optimizer technique to make an appropriate scaling decision to provide the cloud resources for serving of cloud workloads. Besides, we design an extended framework to show interaction between users, cloud providers, and resource provisioning broker in the workload clustering process. The simulation results obtained under real workloads indicate that the proposed approach is efficient in terms of CPU utilization, elasticity, and the response time compared with the other approaches.

Lightweight Security Mechanism over MQTT Protocol for IoT Devices

Security is one of the main concerns with regard to the Internet of Things (IoT) networks. Since most IoT devices are restricted in resource and power consumption, it is not easy to implement robust security mechanisms. There are different methods to secure network communications; however, they are not applicable to IoT devices. In addition, most authentication methods use certificates in which signing and verifying certificates need more computation and power. The main objective of this paper is to propose a lightweight authentication and encryption mechanism for IoT constrained devices. This mechanism uses ECDHE-PSK which is the Transport Layer Security (TLS) authentication algorithm over Message Queuing Telemetry Transport (MQTT) Protocol. This authentication algorithm provides a Perfect Forward Secrecy (PFS) feature that makes an improvement in security. It is the first time that this TLS authentication algorithm is implemented and evaluated over the MQTT protocol for IoT devices. To evaluate resource consumption of the proposed security mechanism, it was compared with the default security mechanism of the MQTT protocol and the ECDHE-ECDSA that is a certificate-based authentication algorithm. They were evaluated in terms of CPU utilization, execution time, bandwidth, and power consumption. The results show that the proposed security mechanism outperforms the ECDHE-ECDSA in all tests.

TAR Based Hotspot Prediction in Cloud Data Centres

In this article, the authors studied hotspots in cloud data centers, which are caused due to a lack of resources to satisfy the peak immediate requests from clients. The nature of resource utilization in cloud data centers are totally dynamic in context and may lead to hotspots. Hotspots are unfavorable situations which cause SLA violations in some scenarios. Here they use trend aware regression (TAR) methods as a load prediction model and perform linear regression analysis to detect the formation of hotspots in physical servers of cloud data centers. This prediction model provides an alarm period for the cloud administrators either to provide enough resources to avoid hotspot situations or perform interference aware virtual machine migration to balance the load on servers. Here they analyzed the physical server resource utilization model in terms of CPU utilization, memory utilization and network bandwidth utilization. In the TAR model, the authors consider the degree of variation between the current points in the prediction window to forecast the future points. The TAR model provides accurate results in its predictions.

Performance optimization of task intensive real time applications on multicore ECUs - a hybrid scheduler

In the current approach of Automotive electronic system design, the multicore processors have prevailed to achieve high computing performance at low thermal dissipation. Multicore processors offer functional parallelism that helps in meeting the safety critical requirements of vehicles. The number of ECUs in high-end cars could be reduced by conglomerating more functions into a multicore ECU. AUTOSAR stack has been designed to support the applications developed for multicore ECUs. The real challenges lie in adapting new design methods while developing sophisticated applications with multicore constraints. It is imperative to utilize the most of multicore computational capability towards enhancing the overall performance of ECUs. In this context the scheduling of the real time multitasking software components by the operating system is one of the key issues to be addressed. In this paper, the state of the art scheduling algorithm is reviewed and its merits and limitations are identified. A hybrid scheduler has been proposed, tested and compared with the state of the art algorithm that offers better performance in terms of CPU utilization, average response time and deadline missing rate both in normal and high load conditions.

Reducing the effects of DoS attacks in software defined networks using parallel flow installation

Software defined networking (SDN) is becoming more and more popular due to its key features, such as monitoring, fine-grained control, flexibility and scalability. The centralized control of SDN makes it vulnerable to various types of attacks, e.g., flooding, spoofing, and denial of service (DoS). Among these attacks, DoS attack has the most severe impact because it degrades the performance of the SDN by overloading its different components, i.e., controller, switch, and control channel. This impact becomes more prominent in SDNs having fine-grained control over traffic for monitoring and management purposes, where large numbers of flow rules are installed. Existing approaches handle DoS attacks in SDN either by dropping malicious packets or by aggregating flow rules, resulting in a legitimate packet drop or loss of fine-grained control over network traffic. In this paper, a parallel flow installation approach is proposed to reduce the effects of DoS attacks, without losing the monitoring capability and fine-grained control over network traffic. The proposed approach installs flow rules in all switches along the path from the source to the destination on a single request from the source; resulting in a considerable reduction of control channel traffic and controller’s utilization. The proposed approach is evaluated by comparing it with the basic SDN controller. The simulation results show that the proposed approach increases the SDN performance in terms of CPU utilization, response time, flow requests, and control channel bandwidth.

Emerging Trends, Techniques and Open Issues of Containerization: A Review

Containerization is revolutionizing the way that many industries operate, provisioning major impact to modern computing technologies because it is extra lightweight, highly portable, energy, resource and storage efficient, cost-effective, performance efficient, and extremely quick during boot up. These often facilitate efficient load balancing, low-level system maintenance, server consolidation (for efficient energy and resource utilization) and replication of instances over geographical locations for better fault tolerance to escalate application reliability. However, some recent literature have addressed various challenges (such as complex networking, persistent storage facilities, cross data centers and multicloud supports, security issues, and lack of available, capable container management APIs, etc.) regarding successful container adoption in industries, which might have resulted in a seemingly meager increase in industrial deployments of containerization over the past few years despite bestowing efficient lightweight virtualization. Moreover, a comprehensive overview of containerizations along with their popularity dynamics has still not been found in contemporary literature, which further extends knowledge gap between developers and available technologies. Hence, current study touches upon different technicalities involved in containerization with potential problems and possible solutions along with various important industrial applications to manifest its existing supports and technical hardships. Finally, we have conducted a comprehensive experimental study to compare the performance of VMs, containers and unikernels in terms of CPU utilization, memory footprints, network bandwidth, execution time and technological maturity using standard benchmarks and observed containers to deliver satisfactory performance in almost all aspects, however, are still not free from issues regarding isolation & security, performance stability, lack of available efficient tools for crossplatform support and persistent storage. Unikernels deliver good performance with VM-like isolation but still need to achieve desired technical maturity (in terms of microprocessor stability, process containment, persistent storage, etc.). VMs, on the other hand, are found to provide stable performance throughout, though bigger memory footprints and slower spin up/down remain their biggest weaknesses.

Load Balancing Evaluation Tools for a Private Cloud: A Comparative Study

Cloud computing turns out to be an emerging technology that revolutionized the world of IT infrastructure. However, since the number of users is increasing daily, the demand for cloud services is increasing too. Thus, congestion occurs on the servers that provide services in the cloud. To avoid congestion, we used load balancer tools such as HAProxy and Nginx to intercept the requests of users and distribute them evenly to the servers. Jmeter is used to measure the performance metrics of least connection algorithm in terms of CPU utilization, response time, and concurrency level. Results showed high performance of HAProxy compared to Nginx in terms of response time and treating requests. Furthermore, we examined the characteristic of availability of the load balancer through deploying redundant load balancers, and we studied the effect of the failure of the load balancer on the quality of service of the end users. Keepalived is used to ensure a smooth transition between the two load balancers. According to the concurrency level, results proved that the number of unsuccessful requests during the failure of the master load balancer is proportionally minuscule compared to the total number of requests sent in a normal situation.

Dataset Popularity Prediction for Caching of CMS Big Data

The Compact Muon Solenoid (CMS) experiment at the European Organization for Nuclear Research (CERN) deploys its data collections, simulation and analysis activities on a distributed computing infrastructure involving more than 70 sites worldwide. The historical usage data recorded by this large infrastructure is a rich source of information for system tuning and capacity planning. In this paper we investigate how to leverage machine learning on this huge amount of data in order to discover patterns and correlations useful to enhance the overall efficiency of the distributed infrastructure in terms of CPU utilization and task completion time. In particular we propose a scalable pipeline of components built on top of the Spark engine for large-scale data processing, whose goal is collecting from different sites the dataset access logs, organizing them into weekly snapshots, and training, on these snapshots, predictive models able to forecast which datasets will become popular over time. The high accuracy achieved indicates the ability of the learned model to correctly separate popular datasets from unpopular ones. Dataset popularity predictions are then exploited within a novel data caching policy, called PPC (Popularity Prediction Caching). We evaluate the performance of PPC against popular caching policy baselines like LRU (Least Recently Used). The experiments conducted on large traces of real dataset accesses show that PPC outperforms LRU reducing the number of cache misses up to 20% in some sites.

RLPAS: Reinforcement Learning-based Proactive Auto-Scaler for Resource Provisioning in Cloud Environment

Public cloud system offers Infrastructure-as-a-Service (IaaS) to deliver the computational resources on demand. Resource requirements of a cloud environment are always fluctuating because of the dynamic nature of the arriving workload, and traditional reactive scaling techniques are employed to deal with this problem. Automated resource provisioning is an effective methodology for handling workload fluctuations by provisioning the resources on demand. Simple reactive approaches affect the performance of elastic system by over-provisioning the resources that substantially increase the costs whereas under-provisioning leads to starvation. An intelligent resource provisioning mechanism overcomes the stated issues by allocating necessary resources by learning the environment dynamically. In this article, RLPAS (Reinforcement Learning based Proactive Auto-Scaler) algorithm is proposed, and it is based on the existing Reinforcement Learning (RL)-SARSA algorithm that learns the environment in parallel and allocates the resources. The performance of RLPAS algorithm is validated using real workloads, and it outperforms existing auto-scaling approaches in terms of CPU utilization, response time and throughput. Further, it also converges at an optimal time step and proves to be feasible for the extensive range of real cloud applications.

Parallel Multi-Class Contour Preserving Classification

AbstractSerial multi-class contour preserving classification can improve the representation of the contour of the data to improve the levels of classification accuracy for feed-forward neural network (FFNN). The algorithm synthesizes fundamental multi-class outpost vector (FMCOV) and additional multi-class outpost vector (AMCOV) at the decision boundary between consecutive classes of data to narrow the space of data. Both FMCOVs and AMCOVs will assist the FFNN to place the hyper-planes in such a way that can classify the data more accurately. However, the technique was designed to utilize only one processor. As a result, the execution time of the algorithm is significantly long. This article presents an improved version of the serial multi-class contour preserving classification that overcomes its time deficiency by utilizing thread-level parallelism to support parallel computing on multi-processor or multi-core system. The parallel algorithm distributes the data set and the processing of the FMCOV and AMCOV generators to be operated on available threads to increase the CPU utilization and the speedup factors of the FMCOV and AMCOV generators. The technique has been carefully designed to avoid data dependency issue. The experiments were conducted on both synthetic and real-world data sets. The experimental results confirm that the parallel multi-class contour preserving classification clearly outperforms the serial multi-class contour preserving classification in terms of CPU utilization and speedup factor.

Resource Utilization in Cloud Computing using Hybrid Algorithm

Objectives: The proposed algorithm represents a multivariate selection and security mechanism for the data which has to be kept at the cloud server. The paper also focuses on the prioritization of the data so that for a given span of time data with high priority is processed first. Methods/Statistical Analysis: The paper has presented a unique way to select encryption algorithm so that miss-utilization is prevented. The paper has included lattice based encryption NTRU (N-th degree Truncated Polynomial Ring Unit) and key based encryption for the security of data. Findings: The unused data bits have been optimized using GA (Genetic Algorithm) and the results have been calculated in terms of total number of completed jobs with priority and the CPU Utilization made against the Jobs. Application/Improvements: Resources which will be utilized at the cloud server would cost the user with some amount but there is optimization of resources in terms of CPU utilization and job completion.

온디맨드 거버너 정책에 따른 작업 스케줄링 기법의 성능 평가

The ondemand governor used in android-based smartphone platforms is a CPU frequency scaling technique. The ondemand governor sets the CPU operating frequency depending on the CPU utilization rate. Job scheduling affects the CPU utilization rate. The power consumption is proportional to the value of operating frequency. Consequently, CPU frequency scaling and CPU utilization rate have an effect on power consumption in a smartphone. In this paper, we evaluated the performance of job scheduling techniques incorporating the ondemand governor in terms of CPU utilization, power consumption, and job deadline miss ratio.