Malicious uniform resource locators (URLs) are prevalent in cyberattacks, particularly in phishing attempts aimed at stealing sensitive information or distributing malware. Therefore, it is of paramount importance to accurately detect malicious URLs. Prior research has explored the use of deep-learning models to identify malicious URLs, using the segmentation of URL strings into character-level or word-level tokens, and embedding and employing trained models to differentiate between URLs. In this study, a bidirectional encoder representation from a transformers-based (BERT) model was devised to tokenize URL strings, employing its self-attention mechanism to enhance the understanding of correlations among tokens. Subsequently, a classifier was employed to determine whether a given URL was malicious. In evaluating the proposed methods, three different types of public datasets were utilized: a dataset consisting solely of URL strings from Kaggle, a dataset containing only URL features from GitHub, and a dataset including both types of data from the University of New Brunswick, namely, ISCX 2016. The proposed system achieved accuracy rates of 98.78%, 96.71%, and 99.98% on the three datasets, respectively. Additionally, experiments were conducted on two datasets from different domains—the Internet of Things (IoT) and Domain Name System over HTTPS (DoH)—to demonstrate the versatility of the proposed model.

An Approach to Resisting Malformed and Flooding Attacks on SIP Servers

Mobile app ads pose a far greater security threat to users than adverts on computer browsers. This is because app developers must embed a Software Development Kit (SDK), called an ad library or ad lib for short, provided by ad networks (i.e., ad companies) into their app program, and then merge and compile it into an Android PacKage (APK) execution file. The ad lib thus becomes a part of the entire app, and shares the whole permissions granted to the app. Unfortunately, this also resulted in many security issues, such as ad libs abusing the permissions to collect and leak private data, ad servers redirecting ad requests to download malicious JavaScript from unknown servers to execute it in the background of the mobile operating system without the user’s consent. The more well-known an embedded ad lib, the safer the app may be, and vice versa. Importantly, while decompiling an APK to inspect its source code may not identify the ad lib(s), executing the app on a simulator can reveal the network behavior of the embedded ad lib(s). Ad libs exhibit different behavior patterns when communicating with ad servers. This study uses a dynamic analysis method to inspect an executing app, and plots the ad lib behavior patterns related to the advertisement into a graph. It is then determined whether or not the ad lib is from a trusted ad network using comparisons of graph similarities.