CPU Usage Research Articles

Analisis Karakteristik Antivirus Berdasarkan Aktivitas Malware menggunakan Analisis Dinamis

Malware, short for “Malicious Software”, is a program specifically designed to perform an activity that can harm software on a victim's device. Examples of commonly found malware include trojans, ransomware and downloaders. It is important for computer users to recognize and avoid malware when using computer devices. Therefore, computer users can overcome malware attacks by using protection software specifically for computer devices using Antivirus software designed to prevent, find, detect, and remove the types of malware that have been mentioned previously. In this study, the dynamic analysis method is used to determine malware activity by running it and monitoring the activity that occurs. This method is usually used to identify the actions that malware performs when it runs. The results showed that the higher the number of malware activities, the higher the metrics tested on the antivirus, such as CPU, memory, disk, and scan time. Regarding the removable drive protection feature, Avast antivirus is relatively more efficient compared to other antiviruses because it has an average CPU usage, low memory, a fairly high detection rate, and fast scan times. Kaspersky Antivirus is relatively the most effective in detecting malware samples with the highest detection rate of 100%. Meanwhile, the Windows Defender antivirus is relatively the weakest in terms of detection rate because it has the lowest detection rate.

MSigHdp: hierarchical Dirichlet process mixture modeling for mutational signature discovery.

Mutational signatures are characteristic patterns of mutations caused by endogenous or exogenous mutational processes. These signatures can be discovered by analyzing mutations in large sets of samples-usually somatic mutations in tumor samples. Most programs for discovering mutational signatures are based on non-negative matrix factorization (NMF). Alternatively, signatures can be discovered using hierarchical Dirichlet process (HDP) mixture models, an approach that has been less explored. These models assign mutations to clusters and view each cluster as being generated from the signature of a particular mutational process. Here, we describe mSigHdp, an improved approach to using HDP mixture models to discover mutational signatures. We benchmarked mSigHdp and state-of-the-art NMF-based approaches on four realistic synthetic data sets. These data sets encompassed 18 cancer types. In total, they contained 3.5 × 107 single-base-substitution mutations representing 32 signatures and 6.1 × 106 small insertion and deletion mutations representing 13 signatures. For three of the four data sets, mSigHdp had the best positive predictive value for discovering mutational signatures, and for all four data sets, it had the best true positive rate. Its CPU usage was similar to that of the NMF-based approaches. Thus, mSigHdp is an important and practical addition to the set of tools available for discovering mutational signatures.

DESIGN AND DEVELOPMENT OF BACKEND APPLICATION FOR THESIS MANAGEMENT SYSTEM USING MICROSERVICE ARCHITECTURE AND RESTFUL API

A thesis is a scientific work completed by students with the aim of developing the knowledge gained during the lecture period. Students at Universitas Trunojoyo Madura (UTM), Faculty of Engineering, particularly Informatics Engineering, carry out their theses manually and on paper. Thesis Management System (TMS) is software designed to help with the thesis execution process by reducing paper usage and increasing time efficiency. Monolithic system development can disrupt the service process if improvements are being made to the system. Therefore, in this research, a Thesis Management System (TMS) will be built using a microservice approach to make it easier to maintain and develop the system, for example, system scalability. As a means of communication between services, TMS is designed and developed using the REST API. TMS has undergone system performance testing to verify that it performs well under certain conditions. The results show that the number of requests increases the performance response time, CPU usage, and memory consumption, with an average resource usage of each service based on a response time of 61.64 ms, CPU usage of 8.64%, and memory usage of 89.47 Mb. As the number of requests on the service increases, so does resource usage in each service, but this has no effect on device performance because the increase is so low.

MAGNet: Machine Learning Guided Application-Aware Networking for Data Centers

Modern data centers are witnessing fast-growing east-west traffic on their network infrastructure due to the highly distributed data center applications. Motivated by the heterogeneity of such application workloads, we propose in this paper an extensible network management architecture called MAGNET which enables application-aware intra-data center networking. The crux of MAGNET is the smart endpoint residing within end-hosts, which is empowered by machine learning combined with lightweight workload tracing to detect workload identities and enable workload-dependent packet tagging. The centralized management plane interface of MAGNET allows network functions to interpret packet tags and perform application-aware packet processing. We demonstrate the feasibility of the architecture via prototype implementation and extensive use case evaluation. Our experiments show that the smart endpoint can fingerprint many real-world applications with 99% accuracy only at 12% additional CPU, and that application-aware data plane can potentially bring substantial benefits in terms of security (e.g., via identity-based microsegmentation), CPU usage (e.g., for intrusion detection) and network latency (e.g., via TCP stack customization).

Integrated Security Information and Event Management (SIEM) with Intrusion Detection System (IDS) for Live Analysis based on Machine Learning

This research builds Security Information & Event Management (SIEM) based on live analysis using machine learning on Intrusion Detection System (IDS). To implement a live analysis technique on IDS using machine learning that is integrated with SIEM, we need a combined system with many processes and services. All processes and services must be orchestrated and combined into one system to make live analysis work. Selection of the right components to be combined into a system that can build live analysis techniques on IDS using machine learning that are integrated with SIEM is needed. In addition, an open-source system for easy deployment is needed in the industrial application. Therefore, this research tries to build the system using most common open-source components for cyberattack live analysis, detection, and monitoring. This research uses a combination of Elastic (ELK) Stack, Slips, and Zeek IDS to build the system. To ensure that the components selected are correct, robust, and reliable, it is necessary to measure the performance of the combined system. Measurement focuses on measuring the performance of resource consumption (CPU and RAM). The proposed system is testing using Denial of Service (DoS) test with 344.1/sec packet. The performance testing shows Elasticsearch is the most component that uses CPU and RAM consumption with 78% of CPU usage and 2300 Mb of RAM usage. The least CPU and RAM consumption is Zeek with 3.5% CPU usage and 225 Mb RAM usage. The proposed system also worked for detecting DoS attacks on the network.

T-Counter: Trustworthy and Efficient CPU Resource Measurement Using SGX in the Cloud

As cloud services have become popular, and their adoption is growing, consumers are becoming more concerned about the cost of cloud services. Cloud Service Providers (CSPs) generally use a pay-per-use billing scheme in the cloud services model: consumers use resources as they needed and are billed for their resource usage. However, CSPs are untrusted and privileged; they have full control of the entire operating system (OS) and may tamper with bills to cheat consumers. So, how to provide a trusted solution that can keep track of and verify the consumers’ resource usage has been a challenging problem. In this article, we propose a T-Counter framework based on Intel SGX. The T-Counter allows applications to construct a trusted solution to measure its CPU usage by itself in cloud computing. These constructed applications are instrumented with counters in basic blocks and added three components in trusted parts to count instructions and defend against malicious CSPs’ manipulations. We propose two algorithms which selectively instrument counters in the CFG. T-Counter is implemented as an extension of the LLVM framework and integrated with the SGX SDK. Theoretical analyses and evaluations show that T-Counter can effectively measure CPU usage and defend against malicious CSPs’ manipulations.

Distributed Anomaly Detection in Smart Grids: A Federated Learning-Based Approach

The smart grid integrates Information and Communication Technologies (ICT) into the traditional power grid to manage the generation, distribution, and consumption of electrical energy. Despite its many advantages, it faces significant challenges, such as detecting abnormal behaviours in the grid. Identifying anomalous behaviours helps to discover unusual user power consumption, faulty infrastructure, power outages, equipment failures, energy thefts, or cyberattacks. Machine learning (ML)-based techniques on smart meter data has shown remarkable results in anomaly detection. However, traditional ML-based anomaly detection requires smart meters to share local data with a central server, which raises concerns regarding data security and user privacy. Server-based model training faces additional challenges, such as the requirement of centralised computing power, reliable network communication, large bandwidth capacity, and latency issues, all of which affect the real-time anomaly detection performance. Motivated by these concerns, we propose a Federated Learning (FL)-based smart grid anomaly detection scheme where ML models are trained locally in smart meters without sharing data with a central server, thus ensuring user privacy. In the proposed approach, a global model is downloaded from the server to smart meters for on-device training. After local training, local model parameters are sent to the server to improve the global model. We secure the model parameter updates from adversaries using the SSL/TLS protocol. Using standard datasets, we investigate the anomaly detection performance of federated learning and observe that FL models achieve anomaly detection performance comparable to centralised ML models while ensuring user privacy. Further, our study shows that the proposed FL-based models perform efficiently in terms of memory, CPU usage, bandwidth and power consumption at edge devices and are suitable for implementation in resource-constrained environments, such as smart meters, for anomaly detection.

Smart Analysis and Detection System for New Host-Based Cryptojacking Malware Dataset

Cryptocurrency is a quickly growing technology in the finance industry, with the first cryptocurrency, Bitcoin, being created in 2009. Each cryptocurrency has its own unique hash value, and cryptocurrency mining involves participating in a guessing competition to release a unique hash into circulation, with the winner receiving a modest bonus in the form of bitcoin. However, as more bitcoins are discovered, it becomes increasingly difficult to obtain more, resulting in a need for extra computer resources and power. Consequently, the increasing popularity of cryptocurrency has led to a rise in cryptojacking malware, which secretly uses victims' computing resources to mine cryptocurrency. This malware can be either web-based or host-based, with similar execution and goals but differing in implementation and injection. Cryptojacking has affected numerous devices worldwide, but few studies have been carried out to detect it, especially the host-based type. Furthermore, the current studies on cryptojacking have limited datasets, which are often outdated or small, and the prediction models developed from these datasets may not be accurate. To address this gap, we conducted a thorough analysis of cryptojacking's behavior, lifecycle, impact, implementations, and possible detection methods. Additionally, we created an up-to-date dataset consisting of 114,985 samples, with 57,948 categorized as benign and 57,037 as cryptojacking. The dataset was used to build a smart cryptojacking detection system, with 5 different convolutional neural network models trained and evaluated against a subset of the dataset. The best performing model achieved an accuracy of 98.4%, an F1-Score of 98.3%, a precision of 98.4%, and a recall of 98.4%. Our proposed method, which involves running Windows executables in an isolated environment and closely monitoring their CPU usage, provides a thorough understanding of cryptojacking malware behavior and enables detection of the malware. The comprehensive dataset collected facilitates efficient detection model development. Additionally, evaluating the dataset with 5 different CNN algorithms and assessing their performance using established evaluation metrics ensures the effectiveness of our proposed method and dataset.

Security Implementation on IoT using CoAP and Elliptical Curve Cryptography

IoT devices typically encompass objects or devices equipped with software and internet connectivity, allowing them to collect data. They have limited processing power, memory, and storage. These limitations can make it difficult to secure these devices, as traditional security measures may not be practical. Implementing security for IoT devices can be challenging due to these characteristics, but several steps can be taken to mitigate these challenges. Elliptic Curve Cryptography (ECC) can secure resource constrained IoT devices by providing a more efficient method of encryption and authentication than traditional methods such as RSA. In this paper, we analyzed the lightweight cryptographic algorithm ECC for securing resource-constrained devices such as IoT and evaluated its performance compared to the RSA cryptosystem. The analysis result indicates that ECC is considered to provide stronger security than RSA for the same key length. The CoAP protocol is applied to devices with limited resources, and an examination of its impact is conducted, considering three key performance metrics: CPU usage, bandwidth efficiency, and message communication latency.

Genetic algorithm with self adaptive immigrants for effective virtual machine placement in cloud environment

In cloud environments, optimization of resource utilizations is one among the predominant challenges. The two sub-research topics are cloud resource prediction and allocation. A few contributions to virtual machine (VM) placement techniques have been identified in the literature. In order to efficiently put up the virtual machine (VM) on the physical machine (PM), a Self Adaptive Immigrants with Genetic Algorithm (SAI-GA) is presented in this study. Based on CPU and memory usage, the proposed technique would forecast the best PM for each VM. The algorithm will adjust itself with the appropriate immigrant based on the history of past VM placement to find the best VM placement. In this paper, the VM live dataset from the CSAP lab at SNU in Korea has been used. For the purpose of demonstrating the significance of the findings, a number of non-parametric tests were used to evaluate how well the proposed SAI-GA performed. The outcomes demonstrate that the suggested approach makes a considerable contribution to the placement of VMs in cloud environments.

Food Image Detection System and Calorie Content Estimation Using Yolo to Control Calorie Intake in the Body

Excess calories in the body can cause obesity and several degenerative diseases, such as diabetes mellitus, heart disease, stroke, hypertension, and others. This system helps people maintain a balanced calorie content that enters the body. The research designs this system using the YOLO algorithm model to detect the type of food which is then developed using the Python programming language to estimate the calories of the detected food. YOLO uses the principle of feature extraction in images that are processed through filters as arrays to perform detection. This system calculates the food calories estimation by multiplying the calories for each food by the amount according to the type of food detected. The calorie value of the food provided is based on the number of calories for each portion of food taken from FatSecret Indonesia. The result is that food detection performance is quite good with average precision, recall, and F1-score values of 0.94, 0.90, and 0.91 respectively, when testing the model. However, when tested on Hugging Face, the performance decreased with the average values of precision, recall, and F1-score respectively, namely 0.84, 0.32, and 0.41. This decrease in performance is because of poor CPU usage and a decrease in image quality when uploaded to the Hugging Face.

A Journey Towards the Most Efficient State Database For Hyperledger Fabric

Hyperledger Fabric is a leading permissioned blockchain platform known for its flexibility and customization. A crucial yet often overlooked component is its state database, which records the current state of blockchain applications. While the platform currently supports LevelDB and CouchDB, this study argues that there is an unmet need for exploring alternative databases to enhance performance and scalability. We evaluate RocksDB, Boltdb, and BadgerDB under various workloads, focusing on memory and CPU utilization. Our findings reveal that each alternative outperforms the existing options: RocksDB excels in throughput and latency, Boltdb minimizes CPU usage, and BadgerDB is most memory-efficient. This research not only provides a roadmap for integrating new state databases into Hyperledger Fabric but also offers critical insights for those aiming to optimize enterprise blockchain systems. The study underscores the significant gains in scalability and performance that can be achieved by reconsidering the choice of state database.

Prediction models for Clustered Virtual Machines in Data Centers

In data center (DC) environments, machine learning algorithms play an important role in resource management to increase efficiency by means of proper predictive monitoring workload trends and adjusting jobs accordingly. In this paper, we propose a system to predict the CPU usage of virtual machines (VMs) of a DC. Our proposal performs clustering of VMs based on their historical information (i.e., time series) by evaluating several traditional ML algorithms using common statistical features of VM time series, which facilitates grouping VMs with similar behaviors and establishing clusters based on these features. Then, training of representative models is performed to finally choose the one with the lowest mean error per cluster. The simulation results show that by performing clustering and training the model with representative time series, it is indeed possible to obtain a low mean error while reducing the local training time per individual VM.

Loop-the-Loops: Fragmented Learning Over Networks for Constrained IoT Devices

In this work, we propose Timed Loop Gears (TLG), as a distributed method for enabling <i>fragmented learning</i> in Resource-Constrained networked IoT edge devices. TLG identifies atomic operations (gears), such as feed-forward and back-propagation, necessary for training Machine Learning (ML) models. Each of these gears executes on a Fog Node (FN) exclusively for each data point at a time rather than the whole dataset in its entirety. Additionally, the networked Edge Devices (EDs) offload the training data to the fog layer using the Message Queuing Telemetry Transport (MQTT) protocol such that the participating FNs subscribe to incoming training data and store them based on topics, simplifying data sharing. TLG enables the FN to then transfer the partially learned weights to the next suitable FN for further training. This looping of weights is repeated across FNs until the training is complete. Through extensive analysis, we observe that, compared to existing distributed ML training approaches, for <inline-formula><tex-math notation="LaTeX">$n$</tex-math></inline-formula> devices, TLG reduces the probability of disruption due to device failure by <inline-formula><tex-math notation="LaTeX">$n^{2}$</tex-math></inline-formula> times. Implementation results of our fragmented learning method demonstrate that, although TLG negligibly increases the memory consumption of the IoT devices by <inline-formula><tex-math notation="LaTeX">$0.8\%$</tex-math></inline-formula> , it reduces CPU usage by almost <inline-formula><tex-math notation="LaTeX">$90\%$</tex-math></inline-formula> . The proposed method proves beneficial for developing and hosting ML models, even on constrained IoT devices, in contrast to existing lightweight ML methods.

Solving the 0/1 Knapsack Problem Using Metaheuristic and Neural Networks for the Virtual Machine Placement Process in Cloud Computing Environment

Virtual machine placement (VMP) is carried out during virtual machine migration to choose the best physical computer to host the virtual machines. It is a crucial task in cloud computing. It directly affects data center performance, resource utilization, and power consumption, and it can help cloud providers save money on data center maintenance. To optimize various characteristics that affect data centers, VMs, and their runs, numerous VMP strategies have been developed in the cloud computing environment. This paper aims to compare the accuracy and efficiency of nine distinct strategies for treating the VMP as a knapsack problem. In the numerical analysis, we test out various conditions to determine how well the system works. We first illustrate the rate of convergence for algorithms, then the rate of execution time growth for a given number of virtual machines, and lastly the rate of development of CPU usage rate supplied by the nine methods throughout the three analyzed conditions. The obtained results reveal that the neural network algorithm performs better than the other eight approaches. The model performed well, as shown by its ability to provide near‐optimal solutions to test cases.

Comparative Study of Machine Learning Models for Bee Colony Acoustic Pattern Classification on Low Computational Resources.

In precision beekeeping, the automatic recognition of colony states to assess the health status of bee colonies with dedicated hardware is an important challenge for researchers, and the use of machine learning (ML) models to predict acoustic patterns has increased attention. In this work, five classification ML algorithms were compared to find a model with the best performance and the lowest computational cost for identifying colony states by analyzing acoustic patterns. Several metrics were computed to evaluate the performance of the models, and the code execution time was measured (in the training and testing process) as a CPU usage measure. Furthermore, a simple and efficient methodology for dataset prepossessing is presented; this allows the possibility to train and test the models in very short times on limited resources hardware, such as the Raspberry Pi computer, moreover, achieving a high classification performance (above 95%) in all the ML models. The aim is to reduce power consumption and improves the battery life on a monitor system for automatic recognition of bee colony states.

Evaluating Web Scraping Performance Using XPath, CSS Selector, Regular Expression, and HTML DOM With Multiprocessing Technical Applications

Data collection has become a necessity today, especially since many sources of data on the internet can be used for various needs. The main activity in data collection is collecting quality information that can be analyzed and used to support decisions or provide evidence. The process of retrieving data from the internet is also known as web scraping. There are various methods of web scraping that are commonly used. The amount of data scattered on the internet will be quite time-consuming if the web scraping is done on a large scale. By applying the parallel concept, the multi-processing approach can help complete a job. This study aimed to determine the performance of the web scraping method with the application of multi-processing. Testing is done by doing the process of scraping data from a predetermined target web. Four web scraping methods: CSS Selector, HTML DOM, Regex, and XPath, were selected to be used in the experiment measured based on the parameters of CPU usage, memory usage, execution time, and bandwidth usage. Based on experimental data, the Regex method has the least CPU and memory usage compared to other methods. While XPath requires the least time compared to other methods. The CSS Selector method is the smallest in terms of bandwidth usage compared to other methods. The application of multi-processing techniques to each web scraping method is proven to save memory usage, reduce execution time and reduce bandwidth usage compared to only using single processing.

베이스보드 매니지먼트 컨트롤러를 위한 부팅 과정 프로파일링 도구

Baseboard Management Controller(BMC) supports server monitoring, maintenance, and control functions using various communication interfaces. However, if an unexpected problem occurs during the device driver initialization process, the BMC may not operate normally. Therefore, a boot process profiling tool that accurately analyzes the device driver initialization process and provides a function to check the analysis result is essential. Existing boot process profiling tools do not specifically provide the device driver initialization process and results required for BMC boot process analysis, forcing developers to use a combination of tools to analyze the boot process in detail. In this paper, we propose an integrated profiling tool for BMC's booting process. The proposed tool provides device driver initialization process analysis, CPU and memory usage analysis, and kernel version management functions. Users can easily analyze the booting process using the proposed tool, and the analysis result can be used to shorten the booting time. Also, the proposed tool is implemented in Linux-based BMC, and it is shown that the proposed tool is more efficient than the existing profiling tool.

Centralized vs. Decentralized: Performance Comparison between BigchainDB and Amazon QLDB

Decentralized databases have gained popularity in the last few years in different areas, such as: traceability, supply chains or finance. Leveraging this type of emerging technology will improve knowledge sharing, as well as the transparency and traceability of the data for digital systems. In a similar way, the characteristics are advertised by the centralized ledger technologies, which are manufactured by large cloud service providers such as Amazon. The present study analyzes the performance of two ledger technologies: BigchainDB (i.e., the decentralized blockchain database) and Amazon QLDB (i.e., the centralized ledger database with transparent and immutable characteristics). For the purposes of comparison, we have integrated these technologies into our traceability platform, which is called the Smart Tracking Platform (STP), and performed a series of experiments enabling us to acquire data for different metrics, such as CPU or memory usage for both the reading and writing operations. The findings of the present study show that QLDB has an overall better performance compared to BigchainDB, based on the metrics that have been considered. From the perspective of database ledger implementation, Amazon QLDB proved to be an integrated solution, easier to use, while BigchainDB comprises a more complex system to be implemented and developed, but is more flexible. Although both systems are almost ready to use solutions for local environments, when it comes to configuration and setting up the communication between nodes within a production environment, BigchainDB adds a layer of complexity from a DevOps perspective, while Amazon QLDB completely overcomes it. Depending on the area considered and the identified needs, both BigchainDB and Amazon QLDB can be considered as suitable solutions for a ledger database.

Network Forensic Serangan DoS pada Jaringan Cloud berdasarkan Generic Framework for Network Forensics (GFNF)

The attacks on cloud-based networks have increased and could lead to various disadvantages such as the inaccessibility of services until the loss of user’s trust. Owncloud is one cloud implementation that runs on a network with more than 200 million users. The aims of these researches are to find digital evidence from DoS attacks. Some DoS attacks are SSH brute force, SYN flood, ping of death, and port scanning on the Owncloud network and then finding the digital evidence such as the attacker's IP, time occurred of the attack, types of the attack, also the resource usage of CPU and RAM. This research uses Wireshark and Snort tools to analyze the network and the method of Generic Framework for Network Forensic (GFNF) as a framework during the simulation process until performing the evidence. The simulation was carried out for 1 minute with 30 trials for each attack. The results of this study found the attacker’s IP, time of the attack occurred, types of attack, and also the increase of the resource usage on CPU and RAM when an attack occurred. The found of results digital evidence such as the attacker's IP, the time occurred of attack, and the types of attack were visualized as a table and presented on the ELK Stack dashboard.