The aim of this research is to quickly detect cross-site scripting (XSS) attacks on the internet based on relationship diagram convolutional networks. Based on the principle and attack process of cross-site scripting attacks, domain knowledge is used to build an XSS ontology to conduct high-level modeling of cross-site scripting attacks, obtain data that can reflect XSS attacks, normalize these attack data, extract attack data word vectors, use them as the input of the relationship diagram convolution networks added to the attention mechanism, and learn attack feature word vectors. After further extracting node characteristics through convolution and pooling, all node characteristics are aggregated and fed into the fully connected neural network. XSS vulnerability detection results are obtained through classification of the activation function, and malicious domain name and malicious IP information are combined as supplementary rules to improve the effectiveness of the vulnerability detection in internet cross-site scripting based on the relationship graph convolution network. Experiments show that this method can accurately detect XSS vulnerabilities, provide comprehensive and accurate attack details, and its performance is better than that of the literature method, which is reflected in the higher accuracy, recall, accuracy and F1 value, and the leading area of the ROC curve. Its detection speed is extremely fast, only 0.03 s, and by combining malicious domain name and IP information, the detection efficiency is further improved, realizing rapid response and effectively maintaining Internet security.

In order to reduce the probability of website users being attacked and maintain the safety of website operation, this study proposes an automatic vulnerability detection method of websites based on associated data. We use plug-ins to scan the website in all directions, establish a scanning database, and classify and store the scanned web data. By applying optimized an a priori association rule algorithm, key features are extracted from web scan data, which are then transformed into input samples for a K-means clustering algorithm. The aim is to efficiently extract feature attributes of website vulnerability data and ultimately construct a text vectorized representation of vulnerability data. Convolutional neural networks can automatically detect website vulnerabilities by using the constructed text vector as input. Experimental verification shows that this method demonstrates comprehensive data coverage, efficient processing speed, and high-precision recognition performance. It not only significantly reduces the clustering analysis time, but also ensures the accuracy and timeliness of vulnerability detection.

Person re-identification (Re-ID) is a critical technology in security systems and video surveillance. However, most of the existing methods focused on precise Re-ID, which not only neglect the transmission overheads, computing energy consumption and carbon emissions, but are unsustainable. Furthermore, the personal semantics is usually blurred and distorted in real-world scenarios due to the bird's eye view (BEV) of cameras. Crossillumination and face-coverings also weakened the key personal semantics. Such deficiencies have resulted in a substantial amount of carbon emissions and poor Re-ID performance. To reduce the video transmission overheads, computing energy consumption and carbon emissions yet guaranteeing the accuracy of Re-ID, this paper proposes a novel semantic communication-based lowcarbon sustainable framework (SC-LCSF) for Re-ID. SC-LCSF adopts the semantic encoder based on an enhanced semanticsaware attention mechanism (ESA-SE) to extract the personal semantics. Only semantic information is transmitted at the semantic layer, which is then decoded into personal IDs by the multi-granularity semantic decoder (MG-SD). Two widely used public datasets, Market-1501 and CUHK03, and a newly curated real-world dataset, HZAU-SCUEC01, are used to train SC-LCSF and to evaluate its performance. Experimental results show that compared to the state-of-the-art (SOTA) methods, SC-LCSF achieves the best Rank-1 and mAP accuracy on all the datasets. Furthermore, SC-LCSF has a significant performance enhancement in low-carbon sustainable computing – the transmission data amount, CPU power consumption, CPU temperature, GPU power consumption, GPU temperature and Re-ID delay have a reduction of 96.8%, 39.6%, 27.9%, 40.9%, 29.7% and 76.6%, respectively.