Cloud Assets Research Articles

Scalable Network Intrusion Detection in Cloud Environments through Parallelized Swarm-Optimized Neural Networks

Cloud computing (CC) offers on demand, flexible resources and services over the internet, to secure cloud assets and resources, privacy and security remains a difficult challenge. To overcome this problem, we proposed a Modified Dove Swarm Optimization Based Enhanced Feed Forward Neural Network (MDSO-EFNN) to examine the network traffics flow that targets a cloud environment.NetworkIntrusion detection systems (NIDSs) are crucial in identifying assaults in the cloud environment, which helps to reduce theproblem. In this study, we gather a NSL-KDD network traffic dataset.Secondly, collected data is preprocessed using Z score normalization to clean the data. Thirdly, Continuous wavelet transform (CWT) is employed to extract the unwanted data. Ant colony optimization (ACO) is used to choose the appropriate data. The selected appropriate data is used to test the process using MDSO-EFNN. The simulation findings of the result use a Python tool. As a result, our proposed method achieves significant outcomes with classification of accuracy (95%), precision rate (97%), sensitivity (98%), and specificity (96%)

Scalable Network Intrusion Detection in Cloud Environments through Parallelized Swarm-Optimized Neural Networks

Cloud computing (CC) offers on-demand, flexible resources and services over the internet, to secure cloud assets and resources, privacy and security remain a difficult challenge. To overcome this problem, we proposed a Modified Dove Swarm Optimization Based Enhanced Feed Forward Neural Network (MDSO-EFNN) to examine the network traffic flow that targets a cloud environment. Network Intrusion detection systems (NIDSs) are crucial in identifying assaults in the cloud environment, which helps to reduce the problem. In this study, we gather an NSL-KDD network traffic dataset. Secondly, collected data is preprocessed using Z-Score normalization to clean the data. Thirdly, Continuous wavelet transform (CWT) is employed to extract the unwanted data. Ant colony optimization (ACO) is used to choose the appropriate data. The selected appropriate data is used to test the process using MDSO-EFNN. The simulation findings of the result use a Python tool. As a result, our proposed method achieves significant outcomes with classification of accuracy (95%), precision rate (97%), sensitivity (98%), and specificity (96%).

Automated Dashboard for AWS Services Monitoring

Abstract: Cloud computing has become an integral part of modern business operations, offering flexibility, scalability, and costefficiency. However, the migration to the cloud brings forth a new set of challenges, with security being a top concern. Organizations need to continuously monitor and audit their cloud environments to identify vulnerabilities, detect threats, and ensure compliance with security standards. In today's digital landscape, cloud computing has become the backbone of many organizations, offering scalability, flexibility, and cost-effectiveness. However, with the increasing reliance on cloud services comes the crucial need for robust security measures to protect sensitive data and ensure compliance with industry regulation. Traditional security audits and monitoring processes are often manual, time-consuming, and prone to human error. They lack the agility and real-time insights required to effectively protect cloud assets. To address these limitations, the Automated Reporting and Dashboards for Cloud Security Audits project leverages Artificial Intelligence (AI) and Machine Learning (ML) technologies to revolutionize cloud security management.

Evaluating and Establishing Baseline Security Requirements in Cloud Computing: An Enterprise Risk Management Approach

In today's digital age, businesses use cloud services to boost their operations exponentially. While this trend presents excellent potential, it is marred by significant concerns over the safety of data stored in the cloud. As the demand for cloud computing increases, businesses face issues ranging from data breaches and provider security to service availability and regulatory compliance. It becomes crucial for organizations to enact stringent security protocols not only to protect user information but also to smoothen operational processes, enforce regulatory compliance, and delineate user behavior norms. To fortify cloud assets, businesses must adhere to foundational security guidelines covering confidentiality, risk analysis, e-discovery, business resilience, and third-party contract obligations. Therefore, while the cloud's advantages are manifold, they have intricate challenges. Businesses must prioritize impenetrable security, craft meticulous risk strategies, and fortify against potential threats to harness the full potential of cloud services while preserving their assets and reputation. These results show that most respondents have a confidence level of 7, indicating a relatively high level of trust in the current cloud security measures. Organizations must continuously assess and adapt their cloud security strategies to keep up with the dynamic digital environment and evolving threats for long-term security and operational efficacy.

Enhancing Security in Cloud Computing with Anomaly Detection Using Machine Learning

Cloud computing has become an integral part of modern business operations, offering unprecedented scalability, cost-effectiveness, and agility. However, the widespread adoption of cloud services has also raised significant security concerns. This paper addresses the imperative need for enhancing security in cloud computing environments through the application of anomaly detection techniques powered by machine learning. The ubiquity of cloud computing has ushered in a new era of digital transformation, enabling organizations to streamline operations and achieve unprecedented efficiency. Nevertheless, the dynamic nature of the cloud, coupled with the evolving threat landscape, has exposed organizations to a spectrum of security challenges. These challenges encompass data breaches, insider threats, and vulnerabilities inherent to the shared responsibility model, which necessitates a collaborative approach between cloud service providers (CSPs) and customers. Anomaly detection, a key facet of cloud security, offers a proactive and adaptive defense mechanism against a wide range of security threats. At its core, anomaly detection relies on the establishment of a baseline of normal system behavior. This baseline is constructed by analyzing historical data patterns, allowing machine learning algorithms to distinguish deviations from the expected norm. Such deviations, often indicative of security incidents or vulnerabilities, trigger alerts for timely remediation. This paper delves into the principles of anomaly detection in cloud computing environments. It discusses the shared responsibility model, the evolving threat landscape, and the need for sophisticated security measures beyond traditional tools. Key anomaly detection principles, such as baseline establishment and machine learning model selection, are elucidated. The paper explores various machine learning algorithms suitable for anomaly detection, including k-means clustering, Support Vector Machines (SVMs), and autoencoders, highlighting their unique strengths and applications in cloud security. Enhancing security in cloud computing through anomaly detection powered by machine learning is essential in safeguarding valuable data and maintaining the integrity of cloud environments. By understanding the intricacies of cloud security challenges, embracing anomaly detection principles, and implementing appropriate machine learning algorithms, organizations can proactively protect their cloud assets and fortify their defenses against emerging threats. This paper serves as a comprehensive guide for organizations striving to secure their presence in the cloud while harnessing its transformative potential.

Multi-Objective Path Optimization in Fog Architectures Using the Particle Swarm Optimization Approach.

IoT systems can successfully employ wireless sensor networks (WSNs) for data gathering and fog/edge computing for processing collected data and providing services. The proximity of edge devices to sensors improves latency, whereas cloud assets provide higher computational power when needed. Fog networks include various heterogeneous fog nodes and end-devices, some of which are mobile, such as vehicles, smartwatches, and cell phones, while others are static, such as traffic cameras. Therefore, some nodes in the fog network can be randomly organized, forming a self-organizing ad hoc structure. Moreover, fog nodes can have different resource constraints, such as energy, security, computational power, and latency. Therefore, two major problems arise in fog networks: ensuring optimal service (application) placement and determining the optimal path between the user end-device and the fog node that provides the services. Both problems require a simple and lightweight method that can rapidly identify a good solution using the constrained resources available in the fog nodes. In this paper, a novel two-stage multi-objective path optimization method is proposed that optimizes the data routing path between the end-device and fog node(s). A particle swarm optimization (PSO) method is used to determine the Pareto Frontier of alternative data paths, and then the analytical hierarchy process (AHP) is used to choose the best path alternative according to the application-specific preference matrix. The results show that the proposed method works with a wide range of objective functions that can be easily expanded. Moreover, the proposed method provides a whole set of alternative solutions and evaluates each of them, allowing us to choose the second- or third-best alternative if the first one is not suitable for some reason.

A Review on Cloud Data Security Challenges and existing Countermeasures in Cloud Computing

Cloud computing (CC) is among the most rapidly evolving computer technologies. That is the required accessibility of network assets, mainly information storage with processing authority without the requirement for particular and direct user administration. CC is a collection of public and private data centers that provide a single platform for clients throughout the Internet. The growing volume of personal and sensitive information acquired through supervisory authorities demands the usage of the cloud not just for information storage and for data processing at cloud assets. Nevertheless, due to safety issues raised by recent data leaks, it is recommended that unprotected sensitive data not be sent to public clouds. This document provides a detailed appraisal of the research regarding data protection and privacy problems, data encrypting, and data obfuscation, including remedies for cloud data storage. The most up-to-date technologies and approaches for cloud data security are examined. This research also examines several current strategies for addressing cloud security concerns. The performance of each approach is then compared based on its characteristics, benefits, and shortcomings. Finally, go at a few active cloud storage data security study fields.

Automatic Building of a Powerful IDS for The Cloud Based on Deep Neural Network by Using a Novel Combination of Simulated Annealing Algorithm and Improved Self- Adaptive Genetic Algorithm

Cloud computing (CC) is the fastest-growing data hosting and computational technology that stands today as a satisfactory answer to the problem of data storage and computing. Thereby, most organizations are now migratingtheir services into the cloud due to its appealing features and its tangible advantages. Nevertheless, providing privacy and security to protect cloud assets and resources still a very challenging issue. To address the aboveissues, we propose a smart approach to construct automatically an efficient and effective anomaly network IDS based on Deep Neural Network, by using a novel hybrid optimization framework “ISAGASAA”. ISAGASAA framework combines our new self-adaptive heuristic search algorithm called “Improved Self-Adaptive Genetic Algorithm” (ISAGA) and Simulated Annealing Algorithm (SAA). Our approach consists of using ISAGASAA with the aim of seeking the optimal or near optimal combination of most pertinent values of the parametersincluded in building of DNN based IDS or impacting its performance, which guarantee high detection rate, high accuracy and low false alarm rate. The experimental results turn out the capability of our IDS to uncover intrusionswith high detection accuracy and low false alarm rate, and demonstrate its superiority in comparison with stateof-the-art methods.

Moving Target Defense-Based Denial-of-Service Mitigation in Cloud Environments: A Survey

With the increased frequency and intensity of denial-of-service (DoS) attacks on critical cloud-hosted services, resource adaptation schemes adopted by the cloud service providers (CSPs) need to be intelligent. Specifically, they need to be adaptable to attack behavior and be dynamic to curb resource over-utilization. The concept of moving target defense (MTD) has recently emerged as an effective and agile defense mechanism against DoS attacks that particularly target cloud-hosted applications. However, the existing surveys that seek to explore this space either focus more on MTD for generic cyberattack mitigation or on DoS attack defense on cloud systems. In this survey, we particularly provide an in-depth analysis on how MTD can help recover critical cloud assets in the face of DoS attacks and how emerging programmable technologies such as software-defined networking (SDN) can be leveraged to achieve that goal. Unlike existing surveys, we categorize DoS attacks on cloud platforms based on their working mechanism. We also discuss the non-MTD-based DoS defense strategies for both cloud and non-cloud infrastructures in order to highlight the pros and cons of MTD-based strategies. We introduce MTD working mechanisms and present how existing research is envisioning MTD’s application in mitigating DoS attacks, both with and without SDN. We also take an in-depth look at the testbed implementations and resilience and performance evaluations of MTD approaches. Finally, we articulate the existing challenges in MTD for DoS mitigation in cloud systems and how these challenges are shaping the future research in this domain.

Raising Awareness about Cloud Security in Industry through a Board Game

Today, many products and solutions are provided on the cloud; however, the amount and financial losses due to cloud security incidents illustrate the critical need to do more to protect cloud assets adequately. A gap lies in transferring what cloud and security standards recommend and require to industry practitioners working in the front line. It is of paramount importance to raise awareness about cloud security of these industrial practitioners. Under the guidance of design science paradigm, we introduce a serious game to help participants understand the inherent risks, understand the different roles, and encourage proactive defensive thinking in defending cloud assets. In our game, we designed and implemented an automated evaluator as a novel element. We invite the players to build defense plans and attack plans for which the evaluator calculates success likelihoods. The primary target group is industry practitioners, whereas people with limited background knowledge about cloud security can also participate in and benefit from the game. We design the game and organize several trial runs in an industrial setting. Observations of the trial runs and collected feedback indicate that the game ideas and logic are useful and provide help in raising awareness of cloud security in industry. Our preliminary results share insight into the design of the serious game and are discussed in this paper.

Resource Specific Security Implementation in Network and Cloud System

The current Unified Threat Management (UTM) systems are limited to standard and specific security practices in cloud and network sub system risk evaluations. In this paper, we recommend dynamic security evaluation to counter live threats by analyzing history of attacks and vulnerability score. Our approach introduces Smart Threat Alert admin (STA) into UTM/firewall engine which scans at regular intervals to discover type, target and implications of attacks at all levels. This technique is an enhanced approach of (QUIRC) structure that captures exact threats experienced in the system but fails to discuss the effects of threats surrounded. This limitation is analyzed using cyber security modeling language (CuSeMOL) on features like vulnerability, probability, type and impact of dynamic attacks. This approach is different from threat specific and static asset risk evaluation system conversely, the proposed novel technique STA analyses internet ports, internet protocol addresses, hostnames and other auxiliary port services to countermeasure the security aspects cloud assets. The proposed setup was implemented at university data center firewall with suitable experiments and received satisfying results

Artificial Intelligence based BOT to Manage and Administer Cloud Services

In the era of cloud computing, every company uses cloud technology for its applications and other infrastructure to provide a highly available and easily accessible user experience. While monitoring and managing these assets becomes a hectic work for the IT admins. On which the Level of Effort (LOE) of the resource allocated will be high and the resource must reach different console for different information. Introducing an AI-powered bot which can monitor and manage the cloud assets will reduce the manpower drastically. Most enterprises currently have very rudimentary systems of resource management where someone in the role of an Azure or resource administrator log on to the Admin Portal of their resources and have to apply filters and search through multiple screens to find even the most basic information regarding utilization and cost. This leads to inefficient management of resources and almost leads to overspending in resources that are being underutilized. The implementation of the project will involve creating a cloud services management bot that can be integrated with an enterprise’s collaboration suite as a way to enhance the enterprise’s modern workspace. The bot is to be trained on a set of query data as part of the artificial intelligence process using the natural language processing packages that are included in the Azure Cognitive Services suite. Once queries are processed, the system will connect with the respective endpoints of the Azure Resource Management REST APIs to retrieve relevant resource utilization information and show that to the end-user.

Dynamic provisioning with structure inspired selection and limitation of VMs based cost-time efficient workflow scheduling in the cloud

Workflow scheduling in cloud computing environments is nowadays a hot topic as scientific workflows application are gradually taking advantage of commercial cloud assets. Common users’ quality of service (QoS) requirements are the respect of defined budget and deadline when executing their workflow job. Since execution cost minimization and completion time minimization are contradictory objectives, addressing such issue through trade-off function approaches have proved to be an efficient way. This paper presents the Cost-Time Trade-off efficient Workflow Scheduling with Dynamic provisioning (CTTWSDP) algorithm. CTTWSDP relies on dynamic VMs provisioning with a limited number of leased VMs, and a cost-time trade-off function over heterogeneous instances to determine the most viable schedule. CTTWSDP also proposed an improved Implicit Requested Instance Types Range (IRITR) evaluation, which is a scheduling concept introduced in our previous work. The IRITR evaluation aims at determining a range of VMs instance types that best suits the workflow execution, in order to avoid overbidding and underbidding that may lead to budget and deadline violation respectively. The results of simulations prove the effectiveness of the proposal. CTTWSDP achieves a 17.09–76.06% higher success rate when compared to four state-of-the-art algorithms. Furthermore, ANOVA along with Tukey–Kramer post-hoc tests have been conducted, revealing the effectiveness of CTTWSDP over three of the baseline algorithm, while for the fourth one the outperformance of CTTWSDP is not statistically significant. An analysis of the standard deviation of the success rate proves that CTTWSDP is more stable in its performance no matter the type and the workload of the workflow. With a standard deviation of 6.73, smaller than the ones obtained by the other algorithms from 18.66 to 34.10.

An efficient latency aware resource provisioning in cloud assisted mobile edge framework

Mobile edge computing is developing as an innovative computing paradigm that gives improved practice to mobile users through low latency connections and enlarged computation limits. As the amount of user requests is time- different, while the computation limit of the edge has is constrained, the Cloud Assisted Mobile Edge computing system is acquainted with improving the adaptability of the edge platform. To give ensured administrations at negligible framework latency, the edge resource provisioning and cloud redistributing of the cloud-assisted mobile edge computing structure ought to be wisely planned effectively. This work proposed a latency aware resource provisioning strategy for distributed cloud-assisted mobile edge computing structure. At first, the framework gets SFC requests for Virtual network functions (VNFs) to use both edge and cloud assets. Here, the efficient parameters, for example, execution time and workload of VNFs are evaluated and Fuzzy logic-based auto-scaling is executed for the overloaded VNFs that need more assets because of the progressively expanded measure of the system packets. Subsequently, the SFC requests are scheduled to the cloud-assisted edge network adequately utilizing the Adaptive Grey Wolf Optimization (AGWO) based asset provisioning algorithm. The exploratory outcomes show the superiority of the presented methodology comparing with the existing techniques as far as system cost, arrival rate, and average response time.

SmartX Multi-Sec: A Visibility-Centric Multi-Tiered Security Framework for Multi-Site Cloud-Native Edge Clusters

Recently, to match the emerging demands for multi-site edge clouds, the cloud-based information and communication technology (ICT) infrastructure is rapidly expanding. To protect distributed edge-based cloud assets from networking-based threats by recognizing suspicious traffic, cloud operators should monitor the overall underlying topology to categorize and identify diversified networking packet traffic, flowing through various paths among virtualized and containerized cloud nodes. Perimeter-based networking security, which employs security appliances in fixed locations, cannot address this visibility challenge. As a result, in this paper, we propose the SmartX Multi-tier Security (Multi-Sec) framework, which aims to provide intuitive and systematic visibility for multi-site edge-cloud security. SmartX Multi-Sec abstracts the underlying networking topology among multi-site edge clusters as multiple onion-ring-based tiers of physical, virtualized, and containerized cloud nodes. It also provides collective DevSecOps automation features for monitoring, visualizing, and filtering targeted networking traffic from the respective tiers of the abstracted networking topology. The resulting flow-centric visibility using SmartX Multi-Sec can be featured with extended Berkeley Packet Filter and eXpress Data Path (eBPF/XDP)-leveraged lightweight flow capture and filtering, three-dimensional onion-ring visualization, and automated deployment of DevSecOps functions. By integrating these features, the Proof-of-Concept (PoC)-version of the SmartX Multi-Sec framework is realized to verify the flexible and scalable flow-centric security for multi-site cloud-native edge clouds.

Multi-objective Virtual Machine Selection in Cloud Data Centers Using Optimized Scheduling

In cloud computing, more often times cloud assets are underutilized because of poor allocation of task in virtual machine (VM). There exist inconsistent factors affecting the scheduling tasks to VMs. In this paper, an effective scheduling with multi-objective VM selection in cloud data centers is proposed. The proposed multi-objective VM selection and optimized scheduling is described as follows. Initially the input tasks are gathered in a task queue and tasks computational time and trust parameters are measured in the task manager. Then the tasks are prioritized based on the computed measures. Finally, the tasks are scheduled to the VMs in host manager. Here, multi-objectives are considered for VM selection. The objectives such as power usage, load volume, and resource wastage are evaluated for the VMs and the entropy is calculated for the measured objectives and based on the entropy value krill herd optimization algorithm prioritized tasks are scheduled to the VMs. The experimental results prove that the proposed entropy based krill herd optimization scheduling outperforms the existing general krill herd optimization, cuckoo search optimization, cloud list scheduling, minimum completion cloud, cloud task partitioning scheduling and round robin techniques.

RETRACTED ARTICLE: Quantitative evaluation method of cloud resources based on work scheduling

With the appearance of the period of large information, the customary information preparing model and asset the executive’s model have been not able to fulfill the developing need. In this way, the recreation of the quantitative assessment strategy for cloud asset dependent on MapReduce figuring mode was advanced right now. The quantitative assessment model and record foundation of cloud assets in Hadoop stage and MapReduce processing mode were abridged. Through the quantitative assessment strategy for cloud assets dependent on MapReduce processing mode, the information was gathered and examined. The outcomes show that the quantitative assessment technique for cloud asset dependent on MapReduce processing model has a decent figuring force and the capacity to anticipate the utilization of cloud assets.

Security Validation Model in Cloud Computing Environment

Private, Public cloud or a unified cloud system, client’s absence of a successful secure computable assessment techniques for handling the security circumstance of its own data foundation overall. This paper gives a quantifiable security assessment framework for various mists that can be gotten to by reliable API. The assessment framework incorporates security checking motor, security recuperation motor, secure computable assessment system, graphical presentation segment & so on. Secure assessment system makes out of many assessment components comparing various fields, for example, figuring, stockpiling, organize, support, application security and so forth. Every component is doled out 3 tuples on the liabilities, score & fix strategy. Framework receives "1 vote" system for a field to check its point & includes synopsis as overall score, & to make high security. We implement the computable assessment for various cloud environment clients dependent on the G Cloud phase. It displays active security examining for one or different clouds with pictorial diagrams & clients to adjust arrangement, expand activity & fix liabilities, in order to increase secureness of cloud assets.

State Transition Model (STM): An optimum solution for preventing co-resident DOS attacks in cloud infrastructure

Co-occupant DoS assaults are discovered as most defenseless dangers in relation to distributed computing which is an asset and generally is obligatory in nature. Co-inhabitant DOS assaults would deplete out the cloud assets which impairs the certifiable cloud clients from executing cloud utilization. Consequently aversion and conclusion of the event related to co-inhabitant assaults results in fundamental assignment of cloud form. In our past research strategy, Two-Player Game Approach (TPGA) is acquainted which points that show evasion of co-inhabitant DOS assaults by learning and arranging the Virtual Machine requested from clients on account of the low, medium and high hazard demands. Anyway this strategy decreased in its execution regarding discovery of hazard dimension of client ask for VMs. There is no predefined method for recognizing hazard estimation of VM assets. And furthermore it would be troublesome keep up and refresh the hazard status data of the cloud assets. This is settled in the proposed research strategy by presenting the system called State Transition Model based defense mechanism (STM) or State Observation based Co-Resident DOS Attack Detection (SO-CRDOS-AD). In the proposed research technique at first hazard estimation of each asset dependent on client demands is assessed by utilizing hazard recognition metric. In light of this hazard metric esteem, state estimation of each client ask for as far as VM is refreshed. The states that are considered in this work are Security State, Vulnerability State, Attacked State, Positive State, Negative State, Degenerate State and Failure State. These state estimations of VM assets are refreshed occasionally with the assistance of Markov Chain display. This examination technique is actualized in the CloudSim condition from which it is demonstrated that the proposed research strategy can guarantee the exact recognition of assault status of VM asks for, therefore the security level can be upgraded.

Analytical and stochastic dependability model of cyber and cloud assets for SCADA system of the critical infrastructure

Analytical and stochastic dependability model of cyber and cloud assets for SCADA system of the critical infrastructure