The detection of software vulnerabilities (or vulnerabilities for short) is an important problem that has yet to be tackled, as manifested by the many vulnerabilities reported on a daily basis. This calls for machine learning methods for vulnerability detection. Deep learning is attractive for this purpose because it alleviates the requirement to manually define features. Despite the tremendous success of deep learning in other application domains, its applicability to vulnerability detection is not systematically understood. In order to fill this void, we propose the first systematic framework for using deep learning to detect vulnerabilities in C/C++ programs with source code. The framework, dubbed Syntax-based, Semantics-based, and Vector Representations (SySeVR), focuses on obtaining program representations that can accommodate syntax and semantic information pertinent to vulnerabilities. Our experiments with four software products demonstrate the usefulness of the framework: we detect 15 vulnerabilities that are not reported in the National Vulnerability Database. Among these 15 vulnerabilities, seven are unknown and have been reported to the vendors, and the other eight have been “silently” patched by the vendors when releasing newer versions of the pertinent software products.
23 September 2019 in IEEE Transactions on Dependable and Secure Computing
Fine-grained software vulnerability detection is an important and challenging problem. Ideally, a detection system (or detector) not only should be able to detect whether or not a program contains vulnerabilities, but also should be able to pinpoint the type of a vulnerability in question. Existing vulnerability detection methods based on deep learning can detect the presence of vulnerabilities (i.e., addressing the binary classification or detection problem), but cannot pinpoint types of vulnerabilities (i.e., incapable of addressing multiclass classification). In this paper, we propose the first deep learning-based system for multiclass vulnerability detection, dubbed μVulDeePecker. The key insight underlying μVulDeePecker is the concept of code attention, which can capture information that can help pinpoint types of vulnerabilities, even when the samples are small. For this purpose, we create a dataset from scratch and use it to evaluate the effectiveness of μVulDeePecker. Experimental results show that μVulDeePecker is effective for multiclass vulnerability detection and that accommodating control-dependence (other than data-dependence) can lead to higher detection capabilities.
27 April 2021 in IEEE Transactions on Dependable and Secure Computing
Automatically detecting software vulnerabilities is an important problem that has attracted much attention from the academic research community. However, existing vulnerability detectors still cannot achieve the vulnerability detection capability and the locating precision that would warrant their adoption for real-world use. In this article, we present a vulnerability detector that can simultaneously achieve a high detection capability and a high locating precision, dubbed Vul nerability Dee p learning-based Locator (VulDeeLocator). In the course of designing VulDeeLocator, we encounter difficulties including how to accommodate semantic relations between the definitions of types as well as macros and their uses across files, how to accommodate accurate control flows and variable define-use relations, and how to achieve high locating precision. We solve these difficulties by using two innovative ideas: (i) leveraging intermediate code to accommodate extra semantic information, and (ii) using the notion of granularity refinement to pin down locations of vulnerabilities. When applied to 200 files randomly selected from three real-world software products, VulDeeLocator detects 18 confirmed vulnerabilities (i.e., true-positives). Among them, 16 vulnerabilities correspond to known vulnerabilities; the other two are not reported in the National Vulnerability Database (NVD) but have been “silently” patched by the vendor of Libav when releasing newer versions.
11 July 2019 in IEEE Transactions on Information Forensics and Security
Data, a key asset in our data-driven economy, has fueled the emergence of a new data trading industry. However, there are a number of limitations in conventional data trading platforms due to the existence of dishonest buyer/data broker. To mitigate these limitations, we posit the importance of a data processing-as-a-service model, which complements conventional data hosting/exchange-as-a-service model. Specifically, in this paper, we introduce a secure data trading ecosystem and present a new blockchain-based data trading ecosystem (hereafter referred to as SDTE). In the ecosystem, both data broker and buyer are not able to obtain access to the seller’s raw data, as they are only getting access to the analysis findings that they require. In other words, we reduce the challenge of securing the dataset to the challenge to secure the data processing. We also build a security model to analyze the data trading market, and describe a new set of trading protocols for the entire data trading market. To demonstrate utility, we implement our proposed secure data trading platform (SDTP) on Ethereum & Intels Software Guard Extensions (SGX) and perform an in-depth analysis.
Source code static analysis has been widely used to detect vulnerabilities in the development of software products. The vulnerability patterns purely based on human experts are laborious and error prone, which has motivated the use of machine learning for vulnerability detection. In order to relieve human experts of defining vulnerability rules or features, a recent study shows the feasibility of leveraging deep learning to detect vulnerabilities automatically. However, the impact of different factors on the effectiveness of vulnerability detection is unknown. In this paper, we collect two datasets from the programs involving 126 types of vulnerabilities, on which we conduct the first comparative study to quantitatively evaluate the impact of different factors on the effectiveness of vulnerability detection. Experimental results show that accommodating control dependency can increase the overall effectiveness of vulnerability detection F1-measure by 20.3%; the imbalanced data processing methods are not effective for the dataset we create; bidirectional Recurrent Neural Networks (RNNs) are more effective than unidirectional RNNs and convolutional neural network, which in turn are more effective than multi-layer perception; using the last output corresponding to the time step for the Bidirectional Long Short-Term Memory (BLSTM) can reduce the false negative rate by 2.0% at the price of increasing the false positive rate by 0.5%.
25 August 2016 in IEEE Transactions on Services Computing
The Software-Defined Network (SDN) is a new and promising network architecture. At the same time, SDN will surely become a new target of cyber attackers. In this paper, we point out one critical vulnerability in SDNs, the size of flow table, which is most likely to be attacked. Due to the expensive and power-hungry features of Ternary Content Addressable Memory (TCAM), a flow table usually has a limited size, which can be easily disabled by a flow table overloading attack (a transformed DDoS attack). To provide a security service in SDN, we proposed a QoS-aware mitigation strategy, namely, peer support strategy, which integrates the available idle flow table resource of the whole SDN system to mitigate such an attack on a single switch of the system. We established a practical mathematical model to represent the studied system, and conducted a thorough analysis for the system in various circumstances. Based on our analysis, we found that the proposed strategy can effectively defeat the flow table overloading attacks. Extensive simulations and testbed-based experiments solidly support our claims. Moreover, our work also shed light on the implementation of SDN networks against possible brute-force attacks.