This PDF file contains the front matter associated with SPIE Proceedings Volume 13175, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Syslogs play a crucial role in maintenance and troubleshooting, as they document the operational status and key events within computer systems. However, traditional methods of anomaly detection in Syslog face challenges due to the sheer volume and diversity of logs, making cross-system anomaly detection difficult. To address those challenges, this paper introduces CATL, a pioneering Contrast Adaptive Transfer Learning with Bidirectional Long Short-Term Memory (BiLSTM), which can effectively extract contextual features of the log sequence from both directions. CATL overcomes the difficulties arising from massive, less-correlated logs between different systems by leveraging a combination of labeled data from source and target systems and optimizing the Contrastive Domain Discrepancy (CDD) metric. This allows CATL to accurately model discrepancies within and across log classes, minimizing intra-class domain discrepancy while maximizing inter-class domain discrepancy in log sequence features from different domains to match existing anomaly detection decision boundaries better. Our empirical studies, conducted on prominent benchmarks including HDFS, Hadoop, Thunderbird, BGL, and Spirit, demonstrate that CATL addresses the syntactic diversity of log systems and outperforms existing methods in cross-system anomaly detection.

With the rapid development of in-vehicle network technology, vehicle safety and protection are facing more and more challenges. The vehicle CAN bus is the main network for vehicle internal communication. However, due to its lack of necessary security mechanisms, the vehicle CAN bus is vulnerable to intrusion attacks. Therefore, developing an effective intrusion detection model is crucial to secure vehicle networks. This study proposes a vehicle CAN bus intrusion detection model based on Bayesian network. This model utilizes the probabilistic reasoning of Bayesian networks and the update characteristics of conditional probability, combined with the characteristic attributes of the vehicle CAN bus, to achieve accurate detection of potential intrusion behaviors. By learning historical data, the conditional probability of the Bayesian network can be updated to achieve real-time detection and prediction of intrusion behavior. In order to verify the effectiveness of the model, we used a real vehicle CAN bus data set for experiments. Experimental results show that the intrusion detection model based on Bayesian network has achieved good results in identifying and predicting intrusion behavior of the vehicle CAN bus. Compared with traditional intrusion detection methods, this model can provide higher accuracy and lower false alarm rate, effectively protecting the security of in-vehicle networks.

In modern e-healthcare systems, healthcare providers usually store users' data in cloud servers. Users wish to obtain relevant diagnostic files through data generated by body sensors. We propose an efficient and privacy-preserving Top- k disease matching scheme (called EPTDMS). EPTDMS uses Density-Sensitive Hashing (DSH) to implement fuzzy search in stage one, employs the cosine value to sort the relevant result, and obtains patient diagnostic files. Improvements are made to address the problems of low matching efficiency, high computational overhead, and high communication volume of most privacy-preserving matching schemes. This scheme achieves disease matching with low computation and communication overhead and reduces the average query time.

Multiparty Private Set Intersection (MPSI) protocols afford protection of set data privacy but concurrently introduce substantial computational overhead. While this overhead is tolerable when dealing with modest-sized sets and a limited number of participants, it becomes burdensome as set cardinality and participant count escalate. Consequently, these protocols exhibit constraints regarding set size or participant quantity, thereby diminishing their practical feasibility in scenarios involving extensive sets and numerous participants. To address these challenges, this paper introduces a novel MPSI protocol based on function secret sharing. Our protocol effectively computes the intersection elements while ensuring the confidentiality of set data, rendering it well-suited for scenarios characterized by a considerable number of candidates and large-scale sets. Extensive testing and comparative analysis against other MPSI protocols are conducted to evaluate the proposed protocol's performance and effectiveness.

This paper presents the need for innovative solutions to optimize computational power networks and the application of quantum computing to tackle real-world challenges. Our research addresses this critical issue by developing a robust pre-training scheme that integrates Quantum Unweighted Quadratic Unconstrained Binary Optimization (QUBO) models with quantum-inspired algorithms. This approach aims to enhance the adversarial robustness of CVQKD systems, ensuring their security in the face of sophisticated hacking attempts. Our experimental results demonstrate that the proposed strategy effectively defends against adversarial attacks while maintaining the integrity of secret keys, showcasing the adaptability and efficiency of QUBO models in quantum communication scenarios. This work not only contributes to the broader application of QUBO models in quantum communication but also provides a robust pre-training scheme that can be generalized and transplanted to other machine learning-assisted systems, significantly improving their security in the face of adversarial attacks.

To protect the location privacy of key nodes in transmission routing, we have studied a trajectory privacy preserving method for transmission resources. By anonymizing the gis information of the whole route, that is, anonymizing all the location information including the first and last stations of the route. Firstly, the internal point of the whole optical cable is protected by grid protection method, and then a method based on Bayesian reasoning is proposed to protect the location privacy of the endpoints including starting point and destination. Through Bayesian inference process, it is proved that the starting point of optical cable routing can be protected by shear the point closest to the starting point and the destination, and the destination location protection algorithm can be obtained in the same way. In order to further improve the endpoints protection performance, we divide the day into different time spans according to the scene of optical cable inspection, and then integrate the anonymization process into this time span. By comparing the two endpoints prediction algorithms with Syn_sub and PBT, it is proved that the proposed endpoints protection algorithm is more effective than the classical algorithm.

Deep Neural Networks (DNNs) have been found to be vulnerable to adversarial attacks. Numerous prior studies aim to address vulnerabilities in trained models, either by patching these weaknesses or by introducing measures to make it challenging or resource-intensive to compute adversarial examples exploiting them. In our research, we take a distinctive approach by exploring and developing a "honeypot" (DMZ-enabled) strategy for Deep Neural Networks (DNNs) to safeguard against adversarial attacks, setting our work apart from existing methods in the protection of DNN models. Unlike other methods, instead of modifying the original model, we split this high dimensional space to get DMZ (demilitarized zone). We intentionally keep weaknesses like traditional honeypot do in the classification model that allow adversaries to search for adversarial examples. The adversary's optimization algorithms gravitate towards DMZ, the examples they generate in the feature space will eventually fall into the DMZ. Then, by checking if the input example falls into the DMZ, we can pick out the adversarial example. More specifically, in this paper, we introduce DMZ in DNN classifiers and show an implementation of a DMZ-enabled defense named zHoneypot. We integrate the current SOTA (state-of-the-art) adversarial example defense methods, i.e., mitigation and detection, into a single approach zHoneypot, we show experimentally that DMZ-protected models can proper handling adversarial examples which generated by state-of-the-art white-box attacks (such as FGSM, PGD, CW, OnePixel and DeepFool) with high accuracy, and with negligible impact on normal benign model. In the end, our defense method achieved a score of 97.66 on MNIST, 86.1 on CIFAR10, and 90.94 on SVHN.

In the field of telematics, the nature of traffic data leads to its scattering, involving numerous traffic objects with individual sensitive information. Consequently, sharing a large amount of data becomes difficult, resulting in the existence of "data silos". To address this issue, this paper proposes a scheme that combines coalition learning and blockchain for the purpose of data sharing and privacy protection. The scheme involves modeling multi-source vehicle data using federated learning and storing the trained model parameters and reputation values of participating vehicles on the blockchain. Furthermore, a reputation value calculation method based on the double subjective logic model is proposed, which analyzes the impact of data source quality on the performance of the federated learning algorithm. This calculation method helps in the selection of the client for federated learning, ensuring efficient screening of data sources, improving sharing efficiency, and achieving privacy protection in data sharing. Finally, a simulation analysis is conducted to evaluate the proposed scheme, and the results demonstrate its capability to filter high-quality data sources in real-time dynamic data exchange scenarios of telematics, thereby enhancing the accuracy of federated learning training.

Cache isolation is a highly effective method for defending against cache side-channel attacks. This approach divides the cache into different isolation domains, assigning distinct domains to mutually untrusted processes, preventing processes from sharing the cache across domains. However, existing solutions have certain limitations. Cache partitioning based on ways has a limited number of isolation domains and may not fully meet users' practical needs. Page coloring schemes require proportional allocation of memory and cache, which is inflexible. This paper introduces ICS, a flexible and secure cache isolation solution. ICS supports up to hundreds of isolation domains, with memory allocation independent of the cache. Additionally, domain management is convenient. ICS is a set isolation solution, with its core being SMT. SMT modifies the mapping relationship between memory and LLC, directing the memory of different isolation domains to distinct cache sets. Implemented with a 1MB 16-way LLC, ICS can support a maximum of 512 isolation domains, with a storage overhead of approximately 1.3% and performance loss of around 1%. It represents a cost-effective method for defending against cache side-channel attacks.

The increasingly severe situation of information security puts forward higher requirements for wireless communication transmission. Minimum shift keying (MSK) offers outstanding advantages, including the constant signal envelope, the narrow main lobe of the power spectrum, and strong channel adaptability. It can be used to increase the modulation map diversity and achieve the purpose of waveform encryption of the transmit signal. As a result, this paper proposes the MSK modulator structure hopping (MSKMSH) method for physical layer secure transmission. Firstly, a set of MSK modulator structures with a space size of 16 is constructed according to signal waveform characteristics. Secondly, the pattern set generated by the structure hopping law is analyzed to ensure traceless conversion between different MSK modulators. Finally, a data-assisted synchronization scheme utilizing unique characters is designed to synchronize the legitimate destination and the source with structure hopping. Through simulation experiments and theoretical analysis, it is demonstrated that the MSKMSH method can realize reliable and secure transmission of radio signals.

Malicious code can reflect its malicious behavior through dynamic API sequences and static PE header information, and deep learning algorithms have made progress in such malicious code detection. The article designs a 1D-CNN-BiGRU network model based on one-dimensional convolutional neural networks and bidirectional gated recurrent neural networks. The model takes API call sequences and PE header information as feature inputs and undergoes convolutional computation and recurrent neural network learning training to further learn the features of malicious code. Through the analysis of experimental results, the correctness of the malicious code verification of this model is demonstrated. The detection accuracy of normal samples on dynamic API call sequences is over 97%, and the accuracy on static PE structures is 95.64%. It has good performance in malicious code detection.

As Internet technology is developed and applied, the problems of large amounts of real-time traffic data and many unknown attacks are becoming increasingly serious, and intrusion detection systems have increased in efficiency and effectiveness. In this paper, a real-time traffic intrusion detection method based on Inception-LSTM deep neural network combining CNN and LSTM is proposed for improving label-based intrusion detection performance. Network traffic records are converted into 2D gray scale graphs. It extracts network traffic features using image processing techniques with high generalization ability. Experimental validation is performed on the publicly available CIC-IDS-2017 dataset, and the results show that the proposed Inception-LSTM neural network improves the detection accuracy and F1-score by 0.5% and 0.7%, respectively; the results of the comparison between the detection done on real-time captured traffic data and the network security devices show that the method is effective and feasible.

Federal learning is an effective distributed learning technology that allows machine learning model training while protecting data privacy. However, with the increase of the number of user -side devices, the calculation burden of users in federal learning will increase. Researchers explore the use of dimension reduction technology to reduce the calculation burden required for model training, but this triggers a problem with low accuracy. This article extracts the dimensions of gradient data by improving the main component analysis method to extract the dimensions of gradient data and reduce communication and calculation burden while protecting the privacy of the client. The experimental results of this article show that under large -scale data sets, the method of this article increases the speed of 50%training and reaches 96% accuracy.

Vulnerability identification is a crucial quality assurance step in software engineering, dedicated to discovering and handling potential errors and abnormal behavior in source code. Most vulnerability detection methods are designed for conventional programming languages. With the widespread adoption of low-code development, there is a need for a vulnerability detection method specifically tailored to low-code environments. Thus, we present a robust low-code vulnerability identification model by integrating Convolutional Neural Network Text Classification (TextCNN) and an attention mechanism. The resulting model is capable of recognizing potential irregular patterns in the low code, assisting developers in promptly identifying and addressing potential software defects. It holds significant importance in enhancing the maintainability, stability, and security of the system. Simultaneously, it offers substantial support for the company's software development efforts and mitigates the risk of software defects. The experimental results demonstrate that the method in this paper can achieve accurate low-code vulnerability identification.

The data center has complete equipment, professional management, and a comprehensive application service platform. A good data center security protection system can help prevent dangers and violations that have a serious impact on the business. However, the level of security protection in data centers varies greatly, and the effectiveness of protection lacks means of testing and verification. This article studies the health evaluation method of data center security protection system, proposes a security protection measure evaluation mechanism. The evaluation method provides strong technical support for the effectiveness verification and promotion of existing data center security protection measures.

This paper designs a flow detection system for marine high-pressure pressure reducing valves, focusing on the hardware selection and inspection methods of the system. The detection system software is completed according to the design requirements to achieve the function of the flow detection system.

In order to address the obstacles posed by the growing security issues of the Internet of Things to the development of big data, this paper conducts in-depth research on the defense of the most harmful DDoS attack. This paper designs a deep learning model for DDoS detection - CBAN. The CBAN model integrates technologies such as 1D-CNN, BiGRU, and attention mechanism for structural design. This model can effectively extract spatial and temporal features of network traffic data for efficient detection of potential DDoS attacks. The CBAN model has shown excellent performance on the CIC-DDoS-2019 dataset.

In this paper, we design a Reputation-based Diffusion Routing (RDR) algorithm based on the reputation evaluation model for selective forwarding attacks in wireless sensor network routing. We introduce the reputation evaluation model to gather the monitoring mechanism of neighbor nodes to obtain more comprehensive security performance. The algorithm judges whether the routing is successful or not through the neighbor node monitoring and evaluates the reputation to defend the selective forwarding attack of internal nodes. It can bypass the area of malicious nodes, improve the success rate of data routing, and reduce energy consumption. The experimental results verify that the RDR algorithm can maintain a high transmission rate with a small energy consumption and can effectively defend malicious nodes.

The code repository stores all the code, configuration files, and historical versions of the enterprise, which is an essential tool in the software development process. However, it also brings unprecedented security threats to enterprises, and the traditional network security architecture has been unable to satisfy the high-security and high-reliability management requirements of the code assets of research institutions. To this end, based on the Zero Trust (ZT) security paradigm, we propose a code repository federated autonomous technology, build a decentralized and distributed autonomous security protection mechanism, realize bidirectional identity authentication and dynamic trust evaluation mechanism for the authorization management of code resources, improve the security and protection level of the code repository, and effectively solve the security protection shortcomings of the existing collaborative research and development environment.

To enhance the fire service system, many countries have conducted extensive research on “Smart Firefighting” platform architecture and achieved significant results. However, there are still numerous challenges that persist in the field. To address the challenges related to storage, computing, security, and scalability of large-scale data in the "Smart Firefighting" domain, we have conducted an in-depth study and comparison of existing "Smart Firefighting" platform architectures. Based on the analysis, we propose a novel architectural for the "Smart Firefighting" platform, which combines the strengths of traditional hierarchical architecture with the Cloud-Edge-End architectural, this integration effectively preserves the key features of both architectural styles, and the proposed fusion model is able to satisfy the computation and storage of large amounts of firefighting data while guaranteeing security. This architectural model can serve as a reference for the construction of highly available, high-performance, and more secure and stable "Smart Firefighting" platforms and holds significant importance in accelerating the development and maintenance of "Smart Firefighting" to ensure long-term social stability and security.

The ever-changing cyber threat landscape, characterized by increasingly sophisticated attacks and evolving malware, demands robust and layered security strategies. Traditional firewalls, while effective as a first line of defense, struggle to keep pace with the complexity of modern threats. This necessitates the adoption of Next-Generation Firewalls (NGFWs) as a vital component of a comprehensive defense-in-depth approach. This research delves into the limitations of traditional firewalls and explores the key capabilities offered by NGFWs. It examines how features like application awareness, deep packet inspection, and SSL/TLS decryption contribute to a more secure and resilient network infrastructure. Through in-depth analysis and relevant case studies, this paper aims to demonstrate the critical role of NGFWs in today's dynamic cybersecurity environment. It emphasizes the critical need for organizations to embrace a defense-in-depth approach, with NGFWs serving as the cornerstone for a robust and resilient network infrastructure. By deploying NGFWs and adopting a comprehensive layered security strategy, organizations can significantly improve their cybersecurity posture and effectively protect themselves against the ever-present and evolving threat actors.

The CAN network bus is one of the most widely used buses in the vehicle, but the lack of sufficient security mechanisms allows criminals to maliciously attack the vehicle, which is a very serious security risk. In this paper, an identity-based encryption algorithm is used to design a secure communication protocol for an in-vehicle CAN network. Compared with the method based on digital certificates, it reduces the overhead of verification, management, and storage of digital certificates. The combination of asymmetric key and symmetric key ensures sufficient security and a small-time cost for the data communication phase. In addition, the Hash Message Authentication Code (HMAC) is used to ensure the integrity and authenticity of the data during the transmission phase. Experimental results show that the proposed protocol can resist forgery attacks, tampering attacks, and replay attacks, and meet the real-time requirements of automotive-grade microcontrollers.

Macro work is a significant Online Labor Platforms (OLPs) operation characterized by higher professionalism for service providers. Therefore, the professionalism assessment for providers of macro work is vital for OLPs. However, due to the high ambiguity of textual data, OLPs often overlook them when evaluating the Service Provider Professionalism (SPP) of macro work. Within OLPs, there is a large amount of textual data, which contains information reflecting their professionalism. Hence, this study proposes a method for evaluating the SPP of macro work on OLPs based on text sentiment analysis: (1) Select professional vocabulary related to a specific type of macro work as sentiment words; (2) Collect texts and score their professionalism values; (3) Calculate the sentiment word professionalism value based on the NBSP algorithm - an algorithm that combines the Naive Bayes and Semantic Orientation Pointwise Mutual Information (SO-PMI) algorithms; (4) Calculate the text professionalism value, namely the SPP value. Algorithm validation results show that compared to baseline algorithms, the NBSP algorithm achieves an increase in the accuracy of calculating text professionalism values by 4.45 - 27.75 percent points. To validate this method's effectiveness, this study conducted a comparative experiment on predicting the annual transaction amounts of IT service providers on a certain Chinese OLP under eight main-stream predictive models, incorporating the feature of SPP reduced MSE by 6% - 12%. This study contributes to expanding research in structuring textual data and text sentiment analysis in OLPs and enhances professionalism assessment for service providers of macro work on OLPs.

Border Gateway Protocol (BGP) is the Internet standard inter-domain routing protocol, has become an important infrastructure of the Internet. Due to the limitation of the initial design, BGP prefix hijacking is a kind of BGP security problem that needs to be solved urgently. This paper proposes a BGP prefix hijack detection algorithm based on Multiple Origin AS (MOAS) event feature, discovering and restoring MOAS events from BGP packets, and creating an event matrix for MOAS events based on the changes of source AS observed from different observation points over time. According to the MOAS event matrix, the eigenvalue of the event is calculated from the perspective of the overall probability of the event and the local visibility of the observation point. Finally, the characteristic value of the MOAS event is compared with the exception event threshold to determine whether the event is an exception event, and then the BGP prefix hijacking is found. Compared with the prefix hijacking detection algorithm based on source AS validity, the detection results of this algorithm have higher accuracy. This method is more practical in detecting prefix hijacking according to the characteristics of the event itself.

Reversible data hiding in images has become one of the research hotspots in the field of information hiding. JPEG images are widely used on social networks, and the research on reversible information hiding technology for JPEG images is of great significance. The existing reversible information hiding algorithms for JPEG images generally lack robustness, which affects the practical application of the algorithms. Inspired by the robust reversible algorithm for spatial domain images, this paper introduces a two-stage embedding method into JPEG images and proposes a new reversible data hiding algorithm for JPEG images based on two-stage embedding. First, apply robust watermarking technology to embed watermark information in the image to achieve robustness, and then apply reversible watermarking technology to embed distortion information to realize reversibility. Experiments have shown that, even if the carrier image is attacked, robust watermarks can still be extracted correctly. Although image restoration cannot be completed due to the destruction of reversible watermarks, the visual quality of the restored image can meet the requirements.

Image encryption has become one of the key techniques for information protection in today's context of increasing number and rising importance of digital communication. The purpose of this paper is to explore a creative image encryption method based on Rijndael algorithm and chaos algorithm, which disambiguates the image, hash value calculation. Then encrypted by Rijndael algorithm and generates a secure ciphertext. Compared to traditional encryption, this algorithm greatly improves the security and speed of encryption by combining multiple encryption techniques, and by conducting several encryption and decryption experiments on this algorithm, we prove that this method has a great prospect for security applications.

With the rapid development of computer network technology, the network structure has become increasingly complex, and the number and types of network devices have also shown explosive growth. In order to better understand, manage and optimize the network, visualization of network topology has become an important research field. However, the current visualization framework and tools have some problems in the practical process, such as complicated application process, single display effect and insufficient interaction ability, which can not meet the actual needs of users. In this regard, this paper will take network topology visualization as the research object, and build a network topology visualization system based on HTML5 through Web technology to realize the innovation of network topology management. Practice has proved that the whole system adopts B/S architecture design, and the front end is mainly based on HTML5, CSS and JavaScript, combined with BootStrap layout framework to complete the design and construction. At the same time, the Canvas is used to complete the drawing and rendering of visual graphics, which greatly improves the efficiency of network monitoring, conforms to the expected goal of network management system in displaying topological functions, and has certain practical application value.

Nowadays, privacy data has widely stored and shared through the cloud storage services from numerous users and organizations. However, existing cloud services are concentrating on improving the data storage amount and ignore the privacy data protection for data owners. In order to protect the privacy data, data owners may encrypt the information and then upload to the cloud servers. Cloud server cannot guarantee the encryption data are not at-tacked or modified by malicious attackers. In this paper, we propose a novel symmetric encryption algorithm to encrypt the privacy data for users and generate the checks for cloud servers and guarantee the privacy cannot be modified and lost by system errors. From our extensive experimental, we can conclude that our proposed model can achieve the protection of privacy by simulating the different attacks with reasonable response period. Furthermore, we compare our model with existing encryption methods and evaluate the effectiveness performance.

This paper tackles the challenges inherent in existing Multi-View Stereo (MVS) methods, which often struggle with scenes that have repetitive textures and complex scenarios, leading to reconstructions that lack quality, completeness, and accuracy. To address these issues, we introduce a novel deep learning network, Clo-PatchmatchNet, which leverages context-aware Transformers for enhanced performance. The network's architecture starts with a feature extraction module that processes image features. These features are then input into a learnable Patchmatch algorithm, creating an initial depth map. This map undergoes further refinement to yield the final, detailed depth map. A key innovation in our approach is the integration of a context-aware Transformer block, known as Cloblock, into the feature extraction stage. This allows the network to effectively capture both global contextual information and high-frequency local details, enhancing feature matching across various views. Our experimental evaluations, conducted on the Technical University of Denmark (DTU) dataset, reveal that Clo-PatchmatchNet outperforms the traditional PatchmatchNet by achieving a 2.5% improvement in reconstruction completeness and a 1.2% increase in accuracy, resulting in an overall enhancement of 1.7%. Moreover, when compared to other contemporary methods, our proposed solution demonstrates superior performance in terms of both completeness and overall quality, marking a significant advancement in the field of 3D reconstruction.

Risk management is a critical factor that influences the success or failure of project management in the development of civil aircraft models. At present, research on project management capability assessment of civil aircraft models is mostly based on qualitative analysis and lacks analysis based on quantitative indicators, which leads to certain limitations in capability assessment. This article uses the Delphi method to obtain the weights of 8 indicators that affect project management ability in the form of a questionnaire. The Analytic Hierarchy Process (AHP) based Analytic Hierarchy Process (YAAHP) software is used to model and calculate the level of project management ability, and suggestions for improving project management ability are proposed.

In the "dual carbon" goal background, massive, dispersed, and ubiquitous distributed resources are becoming an important component of China's power system. However, distributed flexible resources faced with security issues, such as data loss, data tampering, and privacy breaches, during the process of participating in power trading by aggregation method. Therefore, this article proposes key privacy protection technologies to support flexible resource hierarchical aggregation transactions. Through fully analyzing the current situation of flexible resource hierarchical aggregation transaction and the existing privacy leakage problems, this paper designs a collaborative interaction process of flexible resource hierarchical aggregation transactions, and implements privacy protection for electricity enterprises participating in aggregation and benefit distribution in a trustworthy and autonomous manner.

During teleoperation tasks such as disinfection and cleaning inside the space station, the operator on the ground controls the robotic arm within the experimental cabinet to maintain continuous contact with areas of different texture properties by operating the handle of a haptic device. Therefore, real-time and stable haptic feedback can help operators distinguish between different contact areas and assist in completing various teleoperation tasks. This article constructs a 2.5D virtual model integrating shape and texture using curved texture images based on a visual approach. It proposes a textured haptic rendering algorithm, which includes two steps of collision detection and collision response, to achieve virtual force computation integrating contact force and texture force. Finally, the real-time performance and stability of the algorithm are evaluated through subjective and objective experiments.

The Internet has provided great convenience for people, but at the same time, there have also been some problems that infringe on people's interests. How to effectively protect intellectual property rights and prevent works from being tampered with without permission is one of the issues that the Internet is concerned about. Information hiding is a method of hiding secret information into a general carrier of information, in order to prevent others from discovering it during transmission. Therefore, this article proposes a method for hiding encrypted images based on the random LSB matching algorithm. The experimental results show that this method can achieve better hiding effects.

Deep neural networks excel in remote sensing image semantic segmentation, but existing methods, despite their sophistication, often focus on channel and spatial dependencies within identical feature maps. This can lead to a uniform treatment of diverse feature maps, hindering information exchange and impacting model efficacy. To address this, we introduce Feature Map Attention, dynamically modulating weights based on interdependencies among various feature maps. This fosters connections and feature fusion, enhancing the model's capability to represent features. Importantly, this improvement comes with minimal additional computational expense. We also incorporate multipath skip connections, efficiently transmitting features at various scales from encoder to decoder, boosting overall model effectiveness. Our FMAMPN, a lightweight neural network, outperforms other state-of-the-art lightweight models across various datasets.

With the increasing complexity of software and the diversification of vulnerability forms, manual vulnerability mining can no longer meet the needs of software vulnerability mining, and automated vulnerability mining methods are becoming increasingly important. Fuzzing is one of the popular automated vulnerability mining techniques, which is widely used in software vulnerability mining due to its ease of deployment and efficiency. However, fuzzing has strong randomness, which leads to the generation of a large number of redundant and invalid inputs during the fuzzing process, wasting program execution time, resulting in low code coverage, and only a small number of inputs can truly trigger program exceptions. Therefore, the research on oriented fuzzing methods is becoming increasingly important. This article proposes a fuzzing method based on suspicious basic blocks, which uses LLVM in the static analysis stage to analyze the target program and identify the code that may have vulnerabilities. In fuzzing, tracking the execution of these codes, recording edge coverage information, prioritizing the selection of seeds that can trigger potential vulnerability areas for testing, and verifying the effectiveness of the proposed method through experiments.

Traditional Trusted Computing is mainly implemented in the form of boards, chips, etc., and the requirements of hardware modification have greatly limited the widely use of trusted computing. To cope with the dilemma, the idea of designing “Trusted Computing on Chips” becomes a trend with the development of the build-in security module in CPUs. The main challenge lies on how to make full use of processor security features and design a dual-computing security system that meets the requirements of Trusted Computing 3.0. At present, the built-in secure and cryptographic units on processors, such as Phytium and Loongson, and the supported Trust Execution Environment have already provided the foundation for endogenous trusted computing. In this paper, we propose TeTPCM: a TEE Based Endogenous Trusted Platform Control Module (TPCM), which builds an endogenous trusted computing architecture by the collaboration of the TEE and the SoC on phytium processor. Experimental analysis shows that compared with the general-purpose trusted computing chip, endogenous trusted TPCM does not need additional hardware, and is characterized by strong capability, high performance, good scalability, and has better application prospects.

Abstract: In the field of signal processing, array signal processing is an important part. As an important research direction in the field of array signal processing, beamforming technology has great significance in many fields. In this paper, based on the advantages of FPGA technology, the beam control and adaptive algorithm are studied and the design scheme is proposed. Finally, the design is verified by simulation on ModelSim and Matlab platforms. The experiment proves that the error of this scheme is small and the expected convergence effect can be achieved.

Neural Radiation Field (NeRF) is driving the development of 3D reconstruction technology. Several NeRF variants have been proposed to improve rendering accuracy and reconstruction speed. One of the most significant variants, TensoRF, uses a 4D tensor to model the radiation field, resulting in improved accuracy and speed. However, reconstruction quality remains limited. This study presents an improved TensoRF that addresses the aforementioned issues by reconstructing its multilayer perceptron network. Increasing the number of neurons in the input and network layers improves the render accuracy. To accelerate the reconstruction speed, we utilized the Nadam optimization algorithm and the RELU6 activation function. Our experiments on various classical datasets demonstrate that the PSNR value of the improved TensoRF is higher than that of the original TensoRF. Additionally, the improved TensoRF has a faster reconstruction speed (≤30min). Finally, we applied the improved TensoRF to a self-made industrial dataset. The results showed better global accuracy and local texture in the reconstructed image.

The load balancing strategy in the traditional network architecture system is limited to the limitations of the closed operating system of the network device and cannot effectively meet various communication environments. In SDN, a new network architecture with open programming, the load balancing system architecture of the SDN data center is designed. The traffic classification module is used to distinguish large and small flows. A load balancing mechanism based on the improved ant colony optimization is proposed for the forwarding of large flows. A weighted ECMP load balancing mechanism is used for the forwarding of small flows. The improved ant colony optimization considers the difference in link load status and optimizes the parameters of the ant colony optimization. Through experimental simulation and comparison with traditional load balancing algorithms ECMP and DLB algorithms, the advantages of the optimized ant colony optimization in terms of delay jitter, average link utilization and other indicators are verified.

Knowledge graphs are widely used in the field of natural language processing applications. In order to study how to use the structural and attribute information of entities for cross language entity alignment, we have successively borrowed the high-speed gate mechanism of the HGCN model and the relationship aware neighborhood matching model of the RNM model. Firstly, using Graph Convolutional Neural Network (GCN) for knowledge graph embedding learning, and then introducing the method of attribute information and highway gates mechanism to jointly embed the structure and attributes for learning. In entity alignment, relationship aware neighborhood matching is used to improve alignment performance. Therefore, this article proposes a research method for entity alignment based on graph convolutional neural networks and attribute information. Experiments were conducted on the publicly available dataset DBP15k, and from the results, it can be seen that Hits@1 The indicators reached 85.24%, 87.26%, and 94.76% respectively, achieving better experimental results.

With the development of digital media technology, the security and copyright protection of media files have become particularly important. This article proposes an encrypted digital watermark technology based on HTML5 Canvas and JavaScript, which can embed and extract digital watermarks from images in the field of front-end development. In addition, this article discusses the necessity of combining encryption technology with digital watermark technology to provide a security protection mechanism for data security. The paper then introduces in detail the process of using HTML5 Canvas and JavaScript to operate image pixels to implement encrypted digital watermark embedding and extraction and achieve covert embedding of data by modifying the Least Significant Bit (LSB) of image pixels. Finally, by comparing the visual differences of the images before and after the experiment and measuring the PSNR value of each group of experimental images, it can be concluded that the encrypted digital watermark technology can effectively hide and transmit encrypted information while maintaining the visual quality of the image and is suitable for copyright protection of images. and covert communication, etc., which have high practical application value in the field of Web front-end.

Deep learning technology has developed rapidly in recent years, and deep learning-based steganography and steganalysis techniques have also achieved fruitful results. In the past few years, the over-expanded structure of steganalyzers based on deep learning has led to huge computational and storage costs. In this article, we propose image steganalysis based on model compression, and apply the model compression method to image steganalysis to reduce the network infrastructure of the existing large-scale over-parameter steganalyzer based on deep learning. We conducted extensive experiments on the BOSSBase+BOWS2 dataset. As can be seen from the experiment, compared with the original steganalysis model, the model structure we proposed can achieve performance with fewer parameters and floating-point operations. This model has better portability and scalability.

With the rapid development and application of deep learning, its dataset size and network model are becoming increasingly large, and distributed model training is becoming increasingly popular. This article proposes a distributed heterogeneous task scheduling and resource allocation algorithm based on deep learning to address issues such as heterogeneity in resource usage, inability to predict task convergence time, communication time bottlenecks, and resource waste caused by static resource allocation during distributed collaborative training. This algorithm achieves dynamic scheduling and resource allocation of heterogeneous tasks and reduces task completion time in clusters. The experiment shows that the algorithm proposed in this article has significant improvements in both task completion time and system duration.

Cloud computing is a significant leap in the development process of the information industry. It integrates computing, storage, network, and other information resources organically, providing more convenient means for deep sharing and utilization, making the acquisition of information resources no longer limited by time and space. Cloud computing platform has become one of the important components of Information infrastructure, providing an important computing foundation for promoting the Digital transformation of the industry. As cloud computing platforms are increasingly accepted, more and more enterprises are migrating applications and Data migration to the cloud to effectively break through the limitations of local computing and storage resources, and make better use of the development trend of the cloud’s powerful analysis and processing capabilities. Therefore, to ensure the security and credibility of cloud platform Data migration, this paper proposes a key allocation method based on the hash function to support the trusted migration of cloud platform data. Experimental results have shown that the cloud platform key allocation scheme proposed in this article can effectively improve the security capabilities of the cloud platform compared to traditional key allocation schemes.

We found that there are existing academic studies on reverse engineering the firmware that implements Bluetooth and USB protocols to study their security. Electronic Control Unit (ECU) firmware implements the Controller Area Network (CAN) protocol, which is commonly used to implement the communication of the vehicle’s internal network. With the development and growth of electric vehicles, the security of the vehicle network is becoming increasingly important. This paper proposes a method to reverse engineer ECU firmware, which can efficiently help us quickly identify the library functions in the firmware and reduce the errors that may occur during reverse engineering.

Static analysis tools are widely used to ensure code quality and security, especially in large software projects. Recently, the advent of Large Language Models (LLM), such as the Generative Pre-trained Transformer (GPT), seems to present a strong ability to handle tasks about static code analysis. This paper aims to answer the question, can large language model replace static analysis tools? We present an extensive evaluation of ChatGPT’s capabilities in identifying and analyzing issues detectable by three well-known Java static analysis tools: PMD, SpotBugs, and SonarQube. Through a series of experiments, we assess the performance of two versions of GPT, GPT-3.5 and GPT-4, across various categories of code issues. We conduct a detailed analysis of the experiment results and discuss the limitation of using ChatGPT to perform as a static analysis tool. The findings during our research suggest that while GPT, especially GPT-4 performs outstanding marks on the dataset we chose, it is improper to fully replace the static code analyzers at the time. Working as the supplementary of static code analyzers can be a nice way to enhance the code quality ensuring projects.

The CVE is a database of cyber security vulnerabilities that describe the characteristics of vulnerabilities. MITRE ATT&CK formalizes attack patterns (including attack tactics and techniques) through abstract theory, and gives corresponding mitigation schemes, so it is significant to identify ATT&CK attack information from CVE. Since there is no link between the two, this paper proposes a Transformer-based mapping scheme to automatically map CVE’s to ATT&CK, and verify it on the CVE dataset, which has a good mapping effect and provides security operators with Intuitive reference.

Airborne display software WCET assessment relies on actual scenario and traditional static analysis and dynamic test methods can not reflect actual scenario requirements. To achieve WCET value closing to actual scenario for resource saving purpose, firstly the typical architecture of airborne display unit is described; secondly, the worst scenario analysis principles for WCET are identified; thirdly, WCET assessment flow and steps are proposed respectively for non-A661 partition and A661 service partition; lastly, the WCET acceptance principles are given in both software level and system level. This research can help to support airborne display software resource assessment and allocation in the early stage of software design.

In order to study the change of wax deposition thickness with time, an improved DGM (1,1) model was established based on the grey theory model. Taking indoor loop experiment and field pipeline data as examples, the weakening buffer operator and translation transform were introduced to improve the smoothness of the original sequence, and the different models were tested and compared. The results show that the average relative error of the traditional DGM (1,1) model is large, and the error of the improved model is greatly reduced after the weakening buffer operator treatment, and the d2 operator is more suitable for predicting the wax thickness of short sequence and wide spacing sequence. Translation transformation of the original sequence can improve the smoothness of the sequence, and different original sequences have selectivity on the offset of translation transformation function. The effect of weakening buffer operator on model improvement is much better than that of translational transformation. It is feasible to use the modified GM (1,1) model to predict the wax thickness of tube wall, and this method has certain popularization value.

This paper addresses the issue of low parallel efficiency resulting from fixed thread allocation in automatic parallelization compilation technology. The authors employ a genetic algorithm to determine the optimal number of threads for individual parallelizable loops. They then utilize iterative compilation techniques to produce the suitable number of threads for each parallelizable loop structure, thereby enhancing the efficiency of automatic parallelization compilation. The proposed method demonstrated an average performance enhancement of 26% across ten benchmarks in the SPEC CPU2006 test suite and an overall performance improvement of 3.7% in the NPB3.4.2 test suite, thereby indicating the viability and efficacy of the approach. The approach outlined in this paper can be utilized as a benchmark for enhancing the effectiveness of automated parallel computing and promoting the progression of automated parallel computing technology.

At the present stage, just-time software defect prediction has garnered significant attention from researchers due to its granularity and immediacy. Primarily utilizing machine learning classifiers, these models are trained on information from code repositories to predict whether future changes may lead to defects. However, a current challenge with these classifiers lies in the vast number of features, leading to decreased prediction efficiency. These features not only impact model performance but can sometimes result in a decline in predictive accuracy. This paper explores a feature selection technique that combines random forests and self-attention to discard less important features without compromising performance. Through this approach, the number of features required for training is significantly reduced, often to less than 50% of the original features. In our study across six software projects, we observed that using feature selection in the KNN model led to a 9% improvement in the F1 metric and a 6% improvement in the AUC metric compared to logistic regression and Bayesian models. Finally, we applied SHAP for interpretability analysis of the model. This research contributes to enhancing the accuracy and efficiency of just-in-time software defect prediction, providing valuable insights for research and practice in related fields.

The single chip microcomputer has many applications in the research and development of intelligent products and scientific and innovative products. How to improve the efficiency of development and design application effect in the design of single chip microcomputer is worth in-depth study. This paper takes the development of scientific and technological innovation products as the starting point, analyzes and studies how to effectively improve the development efficiency of MCU developers by using Proteus, and gives the specific design ideas and methods. Through the design and analysis of four actual MCU cases, the design efficiency and design effect of developers are verified. Through the statistics and analysis of the actual developer's development data, the design ideas and methods constructed in this paper can effectively improve the time and efficiency of single chip development, and have certain advantages and effects in innovative products, which can be widely used in single chip learning and education.

To meet the trust requirements of complex environments such as cloud platforms, physical trust modules are often virtualized into virtual Trusted Platform Module (vTPM) to function as the root of trust for each virtual machine or container. In remote attestation, vTPM utilize hardware-TPM-based certificate chain extension schemes to ensure its trustworthiness. Existing vTPM certificate chain extension schemes must turn to Privacy Certificates Authority (PCA) for new certificate during every attestation, however, result in excessive computational and communication overheads. To address this issue, this paper proposes a vTPM certificate chain extension scheme based on Multi-Certificate Public Key Infrastructure (MCPKI). This scheme allows unsigned sub-certificates to have the same legitimacy and identity as the root certificate through certificates correlation. In this way, vTPM can have multiple certificates with valid identity and only need to apply to PCA once, which reduces the performance loss caused by repetitive validation in traditional certificate chain extension schemes and enhances the efficiency of Privacy CA certificate issuance. Furthermore, certificate generation is handled by the local vTPM, reducing communication overheads. A new trusted third-party Verification Certificate Authority (VCA) is introduced to share the validation workload of Privacy CA and reduce the reliance on Privacy CA in certificate chain extension. The experiment results demonstrate that this scheme outperforms traditional certificate chain extension schemes in scenarios with multiple certificates requirement.

With the implementation of graded diagnosis and treatment medical reform, the outpatient service volume of community health service centers has increased significantly, which has brought opportunities and challenges to community health service centers. Outpatient service is the first key link for health institutions to receive patients and the first portal for patients to get to know health institutions. The quality of patients' medical service in outpatient service is directly related to the first impression of community health service centers in patients' minds. In order to better serve patients, improve the medical environment, improve patients' trust in primary care, enhance the internal strength and competitiveness of hospitals, optimize the medical treatment process, change the management mode, and carry out standardized management is a step that community health service centers must take now. Therefore, developing an electronic medical record management information system suitable for the needs of community health service centers can serve medical staff and patients, and reduce the workload of relevant personnel to provide more quality services for patients.

Trusted Execution Environment (TEE) find extensive applications in various scenarios such as secure cloud services, databases, big data computing, and blockchain. They establish a secure operational zone, often referred to as an Enclave, for applications. The Keystone Trusted Execution Environment offers a range of security primitives, including memory isolation, secure boot, and remote authentication. It stands out as the first open-source framework under the RISC-V architecture designed for constructing customized TEEs and is currently the mainstream TEE in the RISC-V architecture. However, Keystone itself does not provide protection for Enclaves or security monitors against Cache timing attacks. Recent research has demonstrated that malicious actors, leveraging Spectre attacks, can use Cache timing attack techniques to detect Enclave data through shared caches. This paper proposes a Keystone extension design, calling the extended framework Keyson-LLCI (Last Leavel Cache Isolation). Through security analysis, we prove that it can isolate enclaves and operating systems into different cache paths at the LLC level, thus eliminating the influence of cache side channels and increasing the isolation and security of enclaves. We implemented this hardware extension on an FPGA development board and conducted performance testing and IOZone testing. The test results indicate that Keystone-LLCI can effectively defend against Cache timing attacks, with a performance loss of approximately 14%. The latency of read and write operations increased by 12.4% and 14.1%, respectively. It represents an extension design with relatively low performance loss.

As a typical application basic development tool, C language programming has strong foundation and practicability in practical application. For most beginners of programming, how to effectively carry out the introduction and improvement of C language program is particularly important. Based on the above problems, this paper puts forward the methods and ideas of innovative design of C language program based on game development orientation, actively stimulates the enthusiasm and initiative of application learning, and introduces simple and easy small game content links to make learning interesting and practical. Through the innovative design of C language program based on the game development orientation, the application depth and learning efficiency of beginners are improved, and a good effect of innovative design practice is shown.

Accompanied by the rapid development of the intelligent automobile industry, unmanned driving has also become one of the important research fields of intelligent automobiles. As the future direction of automobile development, its advantage not only lies in the fact that it can enhance the traffic flow per unit time, but also can track and optimize the vehicle driving route and road safety, so as to reduce the development of traffic accidents and safety hazards. Based on this, this paper discusses the unmanned path tracking optimization control system based on MPC algorithm, and illustrates the superiority of this method through simulation, so as to provide certain reference for unmanned path tracking optimization.

Nowadays, the scale of software is getting larger and more complex. The forms of vulnerability also tend to be more diversified. Traditional vulnerability detection methods have the disadvantages of high manual participation and weak ability to detect unknown vulnerabilities. It can no longer meet the detection requirements of diversified vulnerabilities. In order to improve the detection effect of unknown vulnerabilities, A large number of machine learning methods have been applied to the field of software vulnerability detection. Because the existing methods have high loss of syntax and semantic information in the process of code representation, Lead to high false alarm rate and false alarm rate. To solve this problem, this paper presents a software vulnerability detection method based on code attribute graph and Bi-LSTM (Long Short-Term Memory), in which abstract syntax tree sequence and control flow graph sequence are extracted from the code attribute graph of function, Reduce the loss of information in the process of code representation, Bi-LSTM is selected to build a feature extraction model, Experimental results show that, compared with the method based on abstract syntax tree, this method can greatly improve the accuracy and recall of vulnerability detection, improve the vulnerability detection effect for real data sets mixed with multiple software source codes, and effectively reduce the false positive rate and false negative rate.

The atmospheric coherence length instrument, as one of the effective means for assessing atmospheric turbulence, plays a crucial role in the fields of optics and astronomical observations. Existing atmospheric coherence length measurement systems require on-site measurements, posing various inconveniences. In response to this issue, a method for network remote control of the atmospheric coherence length instrument is proposed, accompanied by the development of a corresponding remote control network software system. The introduced system incorporates a remote-control method for the atmospheric coherence length instrument based on the 4G network. The device connects to the Internet via a 4G network, facilitating bidirectional communication between the mobile platform and the atmospheric coherence length instrument through a network server. This paper initially elucidates the communication principles and implementation methods of the remote-control system, followed by the design and development of software system modules for both the device and the Android control side. Rigorous testing demonstrates the method's enhanced instrument versatility and adaptability, enabling data measurements even in extreme and challenging environments. Leveraging IoT technology, the system efficiently allows users to remotely control the atmospheric coherence length instrument, significantly reducing manual management costs.

In order to study secure federated learning for resource-constrained devices such as drones to protect user privacy and data security in drone networks, a blockchain-based secure federated learning scheme for drones is proposed. Currently, researchers focus on transferring models for federated learning after local training using drones, but in reality, drones will be limited in accomplishing local training due to their own resource and arithmetic issues. In this paper, the scheme offloads the training task of the UAV to the local server, and the UAV is only responsible for performing model aggregation and delivery. At the same time, a new consensus algorithm PoE (Proof-of-Energy) is proposed to model the energy and evaluate the arithmetic power of drones, which assigns roles to each drone node within the blockchain network and ensures that the drones effectively participate in the federated learning process. Due to the open and transparent nature of the blockchain, ring signatures are used to replace the traditional signatures in order to protect the private information such as the behavior and identity of each node and the content of block transactions. The experimental results show that the proposed model can ensure that UAVs effectively participate in federated learning. In addition, when there is a poisoning sample to disrupt the training process, the accuracy of the global model can be effectively ensured compared to the traditional scheme.

With the rapid advancement of modern processor technologies, the potential threat to processor memory security from external devices necessitates robust memory protection mechanisms. To ensure the stability and security of processors, Input-Output (IO) physical memory protection mechanisms are commonly employed to prevent unauthorized access to processor memory. However, the redundancy in the mapping relationships between modules in traditional IO physical memory protection mechanisms leads to an increase in processor runtime. This paper, focusing on RISC-V processors, presents an efficient memory protection scheme, named L-OPT, by investigating traditional IO physical memory protection approaches. L-OPT optimizes the mapping relationships between internal modules in conventional memory protection schemes. In comparison with the unoptimized state of the processor in memory testing scenarios, L-OPT demonstrates a 112% efficiency improvement, validating its effectiveness in enhancing processor runtime efficiency.

With the popularity of microservices architecture, how to refactor a monolithic system into a microservices architecture has become a challenge. Traditional system refactoring strategies often focus on a single aspect, such as conducting source code analysis or system load analysis in isolation. This approach has a limited perspective and cannot comprehensively integrate the diverse characteristics of the system for refactoring. This paper proposes a microservices partitioning method for monolithic systems that integrates multiple features. Firstly, it integrates source code and runtime system data to comprehensively acquire system characteristics. By calculating the semantic similarity of class documents, the similarity of abstract syntax trees, and the frequency of interaction between classes, three types of weights are generated. Then, by assigning different weights, a comprehensive weight is calculated to construct an undirected weighted graph representing dependency relationships. Finally, the Chinese Whisper clustering algorithm is used to partition monolithic systems, obtaining suitable microservice modules. Experimental results show that this method can help developers better understand and partition the system, achieving a microservices architecture with high cohesion and low coupling.

Traditional terminal devices generally adopt a process oriented integrated software architecture design scheme, with high coupling between software and hardware modules in various business units. As more and more business modules are connected to intelligent terminals, the software complexity of intelligent terminals also increases exponentially. The complexity of processing new business functions is high, and the development of business software can only be carried out by terminal hardware vendors, making it difficult to meet the needs of flexible business adjustments and rapid installation and deployment. This article designs a software APP based method and system for various data collection needs in intelligent terminals. The hardware of intelligent terminals adopts a modular, scalable, and low-power design, which is adaptable to complex operating environments and has high reliability and stability; By using Docker technology to achieve software and hardware decoupling, APPs communicate through the MQTT protocol, solving the problem of data sharing between different containers and multiple applications within the terminal, and improving the data collection performance of intelligent terminals.

With the continuous development of Software-focused equipment, information equipment, as fundamental platform, needs to update functions, even reconfiguration. In the face of the generalization and servicelization, in order to realize the AI and software-define- equipment, systematic software supportability engineering is adopted. For core supporting software, based on software capability basement model, the architectural design details of the next-generation Information equipment are described, software management process is summarized including software grading, management demand and implementing keys. Based on the capability basement model, the defect rate of review is monitored and controlled. Using the process data model and capability model, the defect rate after delivery is predicted, and reliability is analyzed. The method provides key technically feasible research approach and provides case reference for next generation information equipment.

In this paper, a joint angle selection method basedIn this paper, a joint angle selection method based on Conformal Geometric Algebra (CGA) is proposed for the inverse kinematics model of robot. Then a CGA library is developed and used to construct various kinematic models of 6R serial manipulators. The software of spatial 6R serial manipulator kinematic simulation is developed by Qt and OpenGL technology. This software includes the above-mentioned CGA library and kinematic models. It can display the STL format robot model in three dimensions and modify the size of the model and plan its trajectory.

In order to meet the maintenance requirements of harsh environment and unattended equipment sites, it is proposed to use an embedded Web server as the core in the embedded target turntable system and use HTML web pages to design the human-computer interaction interface to facilitate cross-platform control. Deliver equipment software upgrades and equipment status monitoring, complete human-computer interaction, and achieve remote unmanned control. The article gives the key code for the server implementation. Through experiments, it is verified that the design scheme has little dependence on the control terminal platform and is reliable in operation. It is very suitable for promotion and application in monitoring and control and can improve equipment maintenance efficiency.

Microservices are popular because they have the advantages of loose coupling, low cohesion, and small and autonomous compared to monolithic systems. And when it comes to cloud deployment, it also has a natural advantage. As a result, more practitioners today choose to refactor monolithic applications into one or more microservices, each of which contains a set of partitions composed of components that point to some specific function of the original monolith, so that the entire software system can be represented by a graph, each component can be regarded as a node, and the dependencies between components can be regarded as edges between nodes. In recent years, there has been an approach to using graph neural networks (GNN) to help migrate from monoliths to microservices. However, due to the differences in the research field, some developers rely heavily on the source code of the monolithic system as an important basis for migration, but in the software field, the business information of the project also has a strong symbol for different microservices. Therefore, we will use GNN to comprehensively migrate microservices from the perspectives of business information and source code in the project. The findings indicate that our methodology is superior in efficiency compared to the migration of single features extracted solely from source code.

The urban water supply network affects residents' water use experience and the economic benefits of water supply enterprises, and is an important component of the urban underground pipeline network. With the continuous changes in the pipeline network, the urban water supply network has formed a very complex spatial information system. Due to the difficulty in accurately obtaining information about the water supply network and the traditional manual drawing, manual statistics, and manual analysis management mode which cannot handle the gradually complex information of the network, it is currently very difficult to maintain the network. This article analyzes the current situation of China's urban water supply network. Based on the problems of low information query efficiency and errors in network analysis processes, a water supply network WebGIS management system was developed based on database and GIS technology. The system has a total of 17 functions divided into five modules according to function type: display configuration, data queries, information statistics, network analysis, and map printing. The system uses a spatial database to solve the problem of processing complex network information, and its query function can quickly search for network information. By setting conditions, data filtering can be achieved, greatly improving the efficiency of information retrieval. This system uses GIS technology to visualize the water supply network and provides various click events, making the system interactive and improving the efficiency of information retrieval. In terms of spatial analysis, this article found that the water supply network can be represented by a graph, and algorithms from graph theory can be used to search for pipelines and nodes in the network. Therefore, this system has implemented two major functions of connectivity analysis and valve closing analysis in the network analysis module using Dijkstra's algorithm and breadth-first search algorithm respectively. This system can bring great convenience to the daily maintenance of the water supply network and reduce the error rate of maintenance work. Relevant departments of the government can achieve efficient management of the urban water supply network through the system's data query and network analysis functions.