نشریه فناوری اطلاعات و ارتباطات ایران
سال شانزدهم شماره 61 (پاییز و زمستان 1403)

Nowadays, applications generate huge amounts of data, in the range of several terabytes or petabytes. This data is shared among many users around the world. Distributed systems such as grid and cloud provide a suitable platform for these applications, enabling the use of these diverse mass data applications in a distributed manner. In these systems, they use data replication to face performance problems, guarantee service quality, and increase data accessibility. Replication, despite its many advantages, also brings administrative costs. The balance between the consistency cost of replication and the benefits of replication is a hotly debated topic among researchers in this field. Therefore, paying attention to the consistency of replication plays an effective role in the efficiency of these systems. Many strategies have been proposed by researchers in the field of data replication consistency. Each of these strategies try to reduce consistency costs and provide effective solutions in this field by considering various parameters such as read rate, write rate, old data tolerance rate, number of replicas and communication bandwidth in determining the consistency levels of replicas. In this article, we will examine the concepts related to replication and replica consistency and categorize its types and review previous works in this field. The done works have been compared from the perspective of system type, decision parameters, compatibility model and improved parameters. At the end, the open issues in this field are raised.

Belief analysis or the classification of texts based on the feelings and opinions of users on websites and social media helps people, companies and organizations to make important decisions. Belief mining includes a system for analyzing people's opinions and feelings about an entity such as products, people, organizations, according to the opinions, messages and tweets of users in social media. In this article, the belief analysis of Persian texts based on the messages, comments and tweets of users in social media and websites of 4 datasets using two deep learning methods, CNN, LSTM, taking into account the sense of the word, in two poles, positive and negative with intervals. 2- and 2+ are classified. In the proposed method, first the process of data pre-processing based on character to number conversion, removing the list of extra words and multi-word analysis is done, then for belief analysis and classification of Persian texts CNN, LSTM machine learning algorithm with word sense separation (WSD) is used to Recognize the intensity of emotions according to the words. We call the proposed model CNN_WSD and LSTM_WSD. In the proposed method, the Persian Twitter dataset is used for evaluation and then it is compared with other machine learning and deep learning methods, DNN, CNN, LSTM, in the implementation of this method, python software is used. The accuracy rate of the proposed method for LSTM-WSD and CNN-WSD is 95.8 and 94.3%, respectively.

Deep Neural Networks (DNNs) have demonstrated remarkable performance in various application domains, such as computer vision, pattern recognition, and natural language processing. However, deploying these models on edge-computing devices poses a challenge due to their extensive memory requirements and computational complexity. These factors make it difficult to deploy DNNs on low-power and limited-resource devices. One promising technique to address this challenge is quantization, particularly fixed-point quantization. Previous studies have shown that reducing the bit-width of weights and activations, such as to 3 or 4 bits, through fixed-point quantization can preserve the classification accuracy of full-precision neural networks. Despite extensive research on the compression efficiency of fixed-point quantization techniques, their energy efficiency, a critical metric in evaluating embedded systems, has not been thoroughly explored. Therefore, this research aims to assess the energy efficiency of fixed-point quantization techniques while maintaining accuracy. To accomplish this, we present a model and design an architecture for each quantization method. Subsequently, we compare their area and energy efficiency at the same accuracy level. Our experimental results indicate that incorporating scaling factors and offsets into LSQ, a well-known quantization method, improves DNN accuracy by 0.1%. However, this improvement comes at the cost of a 3× decrease in hardware energy efficiency. This research highlights the significance of evaluating fixed-point quantization techniques not only in terms of compression efficiency but also in terms of energy efficiency when applied to edge-computing device.

Software-Defined Networking (SDN) has been recognized as an efficient approach in the field of communication technology, aiming to improve the performance and efficiency of computer networks, thus reducing costs. One of the key challenges in SDN is load balancing among nodes. Solving this challenge leads to improved response time and network performance. Nowadays, various methods have been proposed for load balancing in SDN, but they have not yet reached the ideal state. In this article, a new method is presented to enhance load balancing and reduce response time. This method utilizes multi-objective evolutionary algorithms and fuzzy weighting. In the proposed method, factors such as bandwidth, traffic status, link buffer, and desired router are taken into account, and the best path and router with desired load balancing for information flows are selected with the minimum time. One prominent advantage of this method is the possibility of performing load balancing automatically without the need for human intervention. Experimental results demonstrate that the proposed method shows a significant improvement of approximately 14.8% in response time compared to other methods, while maintaining load balancing in SDNs. By using the proposed method, in addition to improving service quality and user satisfaction, response time will also be enhanced. In summary, the proposed method is introduced as a viable approach in SDNs and exhibits superiority over existing methods.

Fuzzing software is a method for finding security vulnerabilities in applications. In this method, by sending random data to the program, attempts are made to find cases that lead to undesirable behaviors and errors such as memory corruption or unauthorized access. One of the proposed methods for improving and enhancing fuzzing is the use of symbolic analysis and dynamic-symbolic execution. In this method, in addition to generating random data, logical analysis of the program and its symbolic execution are used to generate data that can cover new paths in program execution. In this research, we have shown that the dynamic-symbolic execution method can be used for fuzzing network protocols and also improve this process. For this purpose, the first framework for hybrid fuzzing of network protocols has been designed and implemented. The results on two services dcmtk and dnsmasq show that hybrid fuzzing performs better in terms of code coverage compared to traditional fuzzing. Branch coverage in the dcmtk service improved by 2.71 percent compared to AFLNet, which was able to make the negative performance of NyxNet compared to AFLNet positive. Also, branch coverage in the dnsmasq service improved by 37.72 percent compared to AFLNet and by 11.82 percent compared to NyxNet.

One of the important issues and concerns in agile enterprise architecture is the documenting and modeling of requirements and artifacts. Architectural models and documents should be produced and updated as necessary and while meeting the requirements and concerns of the stakeholders, they should be presented in the shortest time and at the lowest cost. On the other hand, methods and tools should be used that are applicable and simple, and while using agile practices and modeling standards, have the necessary comprehensiveness. Researches has been done regarding the modeling and documenting of agile enterprise architecture. Since there is no comprehensive, standard and minimal method for modeling and documenting agile enterprise architecture in related researches, therefore, in this paper, a method for modeling agile enterprise architecture is presented and evaluated by the combined (qualitative + quantitative) method. Qualitative evaluation is performed through a case study. For quantitative evaluation, the AHP method is used to weighting and ranking the criteria. Evaluation criteria are extracted based on library studies and experts' opinions. In order to quantitatively evaluate the proposed method, six criteria have been proposed in this paper, which are: production of required and necessary artifacts, reduction of modeling time, reduction of cost, improvement of stakeholders satisfaction, improvement of applicability and increase of simplicity, and criteria of production of required and necessary artifacts, has won the highest rank.

. For integrated management and command and control over situations and issuing appropriate orders, it is necessary for an organization like the police to intelligently monitor all issues and create an intelligent command and control center to complete missions efficiently. One of the components of increasing security is people's cooperation with the police in order to scientifically detect crimes, as well as benefiting from the latest technologies at the organization level. Despite the wide role of information technology in Faraja, in order for Faraja to carry out the assigned missions and establish order and security, the police organization must be transformed from a technology-oriented to a smart police, so the purpose of this research is to "provide a model It is a native strategy of the smart police of the Islamic Republic of Iran. This research is applied-developmental in terms of type and descriptive-analytical in terms of method. The statistical population of this research is targeted among the experts whose common feature is management experience in smart projects, with expertise in the fields of systems design or information technology, a strategic job position, which is estimated at 50 people. The analysis of the information and data obtained from the study of the sources has been done in a qualitative way and the information obtained from the questionnaire has been analyzed using the method of structural equations with the partial least squares approach and the use of SmartPLS software.

The possibility of developing intelligent systems compatible with the environment has increased by expanding computer engineering fields. Self-adaptive systems are one of the types of software systems that change their behavior according to system conditions and environmental conditions and adapt themselves to it. Although the evaluation of the performance of these systems has been covered in most researches, the evaluation of their quality has remained closed. Therefore, designing an approach is an essential issue to evaluate the quality of self-adaptive systems. The first challenge, the quality indicators of these systems are not fixed, specific, and definite parameters. For example, one of the qualitative indicators for a self-adaptive system is the ability to run software on different operating systems. This indicator may have different degrees of importance for different experienced individuals. The next challenge, some qualitative indicators are not mathematical variables, but rather a linguistic variable between users and experts, indicating that these qualitative quantities are fuzzy variables and can be completely formulated by fuzzy logic. In this article, a new genetic-fuzzy type 2 approach has been proposed to evaluate the quality of the systems based on their quality indicators. The fuzzy logic type 2 is used to describe qualitative indicators, and a genetic algorithm is utilized to determine the optimal fuzzy weights of the qualitative indicators.

In the proposed method, it has been tried to compare the self-adaptive systems from two dimensions, including the software dimension and their self-adaptive dimension. This is despite the fact that most of the existing research deals with only one dimension. In order to evaluate the proposed method, a traffic control system called InSync, which is an adaptive traffic control system (ATCS) and contains multiple qualitative factors, has been used. The obtained results confirm the effectiveness of the proposed method due to the lack of scenario generation, the ability to extend the method to all software quality parameters, and its simplicity. Also, the method is more comprehensive than other existing evaluation models. In addition, it has been tried to compare the self-adaptive systems from two dimensions, including the software dimension and their self-adaptive dimension. This is despite the fact that most of the existing research deals with only one dimension.

Cloud computing provides computational services such as servers, memory, storage space, databases, networks, software, analytics, and information as virtualized resources through the internet to offer faster innovation, flexible resources, and cost savings at scale. Although cloud computing service providers are innovatively expanding their services, trust is one of the major obstacles to the progress of this matter. Trust is the biggest issue in cloud computing since trust is an effective guarantor during interactions between the users and the providers. Trust is one of the most fundamental methods for increasing confidence in resources provided in the cloud environment and is important in cloud business environments. With the increasing number of cloud services providers in the cloud computing environment and the number of users, the selection of provider has become a major challenge. The Coa algorithm has a higher convergence speed, at least by 5.9%, compared to the studied algorithms. In this research, an optimization approach based on a metaheuristic process using the COA algorithm combined with the K-means clustering algorithm is proposed to solve the optimization problem of selecting the best provider in the trust management third-party component layer based on parameters. In this method, while reducing trust evaluation time, the accuracy of user access to the most trusted provider based on user priorities has increased compared to previous methods. This can increase user confidence and improve the quality of service providers.

The Rhynchophorus ferrugineus is a major pest that serves as a carrier for bacterial and fungal diseases, causing significant damage to palm plantations when observed on farms. Nowadays, advancements in wireless communication environments have made it possible to develop low-cost, energy-efficient, multi-functional, and short-range sensor nodes for tracking this pest in palm plantations. In existing target tracking algorithms, the probability of losing the target increases with its speed. Therefore, this paper proposes a new method for target tracking that reduces the likelihood of losing the target. Additionally, considering the energy constraints of battery-powered sensor nodes, we need a scheduling mechanism for their sleep and wake-up cycles to enhance the network's lifespan. To improve energy consumption, this paper utilizes a time scheduling approach to adjust the sleep and wakeup periods of nodes using the Cat Swarm Optimization algorithm and Fuzzy Logic optimization. By simulating the proposed method and comparing it with the Tracking-45-Degree-vectors method in the Opnet simulator, it can be observed that the proposed protocol performs significantly better. Specifically, the end-to-end delay rate improves by 27.02%, the media access delay rate improves by 2.01%, the throughput rate improves by 0.62%, the signal-to-noise ratio improves by 3.28%, and the average battery energy consumption improves by 8.77% compared to the Tracking-45-Degree-vectors protocol. It is worth mentioning that the proposed algorithm has been simulated and tested for a single target scenario.

Software mutation testing is one of the effective methods to evaluate code quality and detect hidden errors. However, this method faces challenges such as producing equivalent mutants and the process being time-consuming. The test cases designed for software testing must have the necessary sufficiency. For this purpose, the criterion of mutation test score is used. One of the main issues related to software mutation testing is the generation of equivalent mutants. Equivalent mutants are mutants that do not change the behavior of the program and have the same output as the original program. Identifying these mutants can make the mutation testing process time-consuming and costly. It is also possible that these mutants are mistakenly classified as hard to kill mutants. While hard to kill mutants can be rejected by changing the test items and strengthening them. In this paper, we present an efficient method for classifying program mutants to identify and separate equivalent mutations from hard to kill mutants. Using machine learning algorithms, we intend to optimize this process and improve the efficiency and effectiveness of the method. This is done by extracting different features from the mutants and using them to train machine learning models.

Twitter List recommender systems have achieved high prediction accuracy by leveraging diverse user and List information alongside complex algorithms. However, explainability remains a significant challenge in these systems. Providing meaningful explanations along with a set of recommendations can enhance user trust and satisfaction, assisting them in informed decision-making. In this paper, we present a model for the automated generation of personalized descriptions as explanations for recommended Twitter Lists. Specifically, our model selects frequently used hashtags from the content of the recommended List, establishing semantic relationships with the user's activity history. The aim is to present the List's subject in an understandable and personalized manner through a generated description. Through experiments conducted on a real Twitter dataset, our proposed model demonstrates its capability to generate explanations for a high percentage of the recommendations provided by a recommendation model.

Today, social networks have brought a lot of convenience to their users, but there are still problems of not respecting privacy. This has caused the security and privacy protection of social network users to be important. Users share a large amount of personal data on social networks, attackers can obtain sensitive personal information simply by using social networks. and carry out various types of attacks and identity theft. Many social networks do not have their own data centers and usually store user data in third-party data centers. These centers can share this data with others and provide user data to other organizations without their knowledge. Owners of social networks can provide users' data to information agencies or use them for advertising purposes. In this research, in the first part, we have described the introduction of the subject, and then in the second part, we have discussed the literature on the subject, which is mostly focused on the centralized methods of privacy protection, and then in the third part, we have described the research method. And we have presented a method that increases the privacy of users in social networks to a great extent by using blockchain. In the fourth part, we have described the implementation of the research and finally we have concluded and summarized this research.

Security is a main goal in the design of industrial Internet of Things network. Due to the ever-increasing developments in the Internet of Things, it is necessary to use new methods to detect active network attacks. In this article, an intrusion detection system for industrial Internet of Things is proposed. This system uses the combination of gray wolf meta-heuristic algorithms (GWO) and decision tree (DT), nearest neighbor (KNN) and artificial neural network (ANN) classification algorithms. First, the data is pre-processed and then normalized, in the next step, data feature extraction is performed using the gray wolf algorithm to extract its independent and effective features. Then it is trained using classification algorithms and finally evaluated. The obtained results show that the use of the combined GWO-ANN algorithm with 93.22% accuracy has a better performance in detecting attacks. Also, the ANN algorithm is more accurate than the DT and KNN algorithms when combined with the GWO algorithm. ,

Link prediction is an important to check link between nodes in social networks. The modeling of social networks leads to emergence of signed, directed and weighted social networks. The relationships of users in social networks are characterized by subjective, asymmetric and ambiguous aspects related to this domain, then both terms of trust and distrust are challenging. To solve the problem of sparsity in networks and overcome ambiguity in relationships, a trust-distrust method based on fuzzy computational is proposed to calculate strength of links. The purpose of proposed link prediction is to solve problem of sparsity in signed social networks by combining descriptive features of users with the direct influence of top nodes and the indirect influence of common nodes on rating prediction. Trust is determined by a Mamdani fuzzy system based on mirroring of similarity fuzzy features, overall trust and overall distrust. The evaluation of the proposed method was done with the accuracy measure on datasets of Epinions and Slashdot. The accuracy of proposed method in Epinions and Slashdot datasets is 0.991 and 0.998, respectively. The obtained results show that proposed method works well for problem of data sparsity in signed social networks and show the effectiveness of proposed model.

Fault detection in photovoltaic (PV) arrays is necessary to increase the output power and also the useful life of a PV system. The presence of conditions such as partial shade, high impedance faults, and the maximum power point detector (MPPT) system make the fault detection of PV in environmental conditions more challenging. The literature identified and classified defects just in few scenarios. In this study two-dimensional scalograms are generated from PV system data. The VGG16 as a pretrained convolutional neural network is used for feature extraction. Finally, to identify and classify faults in the PV system a fully connected neural network is trained. Unlike the previous methods proposed in the literature on the subject of defect detection and classification, various defective cases with MPPT combination are considered in this research. It has been shown that the proposed method including pre-trained CNN performs better than the existing methods and achieves an error detection accuracy of 83.375%.

In recent years, cyber attacks on the critical infrastructures of countries have increased significantly. The types of critical infrastructure and their dependencies based on national requirements are different from one country to another. Disruption in the mission or services of a critical infrastructure has a cascading effect on other infrastructures and causes them serious problems. In different studies, different approaches have been taken to model these dependencies, the important point is not to generalize those models to other countries due to the national requirements of each country. In this research, by forming 11 focus groups consisting of top and middle managers of each infrastructure area, the network analysis method based on the DEMATEL technique was used, and the most influential and influential critical infrastructure from a cyber perspective on other critical infrastructures was identified and the relationship between the infrastructure Critical issues and their prioritization were determined from a cyber perspective. The results of this research can be useful in the design of the national warning sharing system in order to calculate the national situational awareness in the cyber field and other researches based on the dependence of critical infrastructures.

In order to use the Internet of Things as a secure infrastructure, there are various challenges and problems, of which security is one of the most important. Establishing security in such networks has many complications due to the limitation of various resources, including processing resources and low energy, and there is a need to establish a kind of compromise between security and available resources. These conditions have caused security to become an important challenge in these networks, and various methods have been presented to improve and optimize this challenge. Accordingly, in this article, a lightweight encryption solution based on symmetric and asymmetric encryption is presented to ensure data security on the Internet of Things. In the proposed method, first, the main data is encrypted by the symmetric Bluefish algorithm, and then its key is secured with the help of the elliptic curve encryption algorithm, so that as a result, data security can be ensured in a short time and with high security in infrastructures based on the Internet of Things provided. In the end, the proposed solution has been evaluated through the Eclipse simulator and by testing on the data volume of 20 to 1000 kilobytes. The simulation results show that the proposed method performs more optimally compared to other encryption algorithms in terms of evaluation criteria such as execution time and encryption and decryption throughput. These results indicate that the proposed solution, while establishing security, has had the least negative impact on the processing resources of IoT nodes.