The publish/subscribe model has gained prominence in the Internet of things (IoT) network, and both Message Queue Telemetry Transport (MQTT) and Constrained Application Protocol (CoAP) support it. However, existing coverage-based fuzzers may miss some paths when fuzzing such publish/subscribe protocols, because they implicitly assume that there are only two parties in a protocol, which is not true now since there are three parties, i.e., the publisher, the subscriber and the broker. In this paper, we propose MultiFuzz, a new coverage-based multiparty-protocol fuzzer. First, it embeds multiple-connection information in a single input. Second, it uses a message mutation algorithm to stimulate protocol state transitions, without the need of protocol specifications. Third, it uses a new desockmulti module to feed the network messages into the program under test. desockmulti is similar to desock (Preeny), a tool widely used by the community, but it is specially designed for fuzzing and is 10x faster. We implement MultiFuzz based on AFL, and use it to fuzz two popular projects Eclipse Mosquitto and libCoAP. We reported discovered problems to the projects. In addition, we compare MultiFuzz with AFL and two state-of-the-art fuzzers, MOPT and AFLNET, and find it discovering more paths and crashes.

In this paper, we propose a new attribute-based proxy re-encryption scheme, where a semi-trusted proxy, with some additional information, can transform a ciphertext under a set of attributes into a new ciphertext under another set of attributes on the same message, but not vice versa, furthermore, its security was proved in the standard model based on decisional bilinear Diffie-Hellman assumption. This scheme can be used to realize fine-grained selectively sharing of encrypted data, but the general proxy re-encryption scheme severely can not do it, so the proposed scheme can be thought as an improvement of general traditional proxy re-encryption scheme.

Currently, hidden Markov-based multi-step attack detection models are mainly trained using the unsupervised Baum–Welch algorithm. The Baum–Welch algorithm is sensitive to the initial values of model parameters. However, its training uses random or average parameter initialization methods, which frequently results in the model training into a local optimum, thus, making the model unable to fit the alert logs well and thereby reducing the detection effectiveness of the model. To solve this issue, we propose a pre-training method for multi-step attack detection models based on the high semantic similarity of alerts in the same attack phase. The method first clusters the alerts based on their semantic information and pre-classifies the attack phase to which each alert belongs. Then, the distance of the alert vector to each attack stage is converted into the probability of generating alerts in each attack stage, replacing the initial value of Baum–Welch. The effectiveness of the proposed method is evaluated using the DARPA 2000 dataset, DEFCON21 CTF dataset, and ISCXIDS 2012 dataset. The experimental results show that the hidden Markov multi-step attack detection method based on pre-training of the proposed model parameters had higher detection accuracy than the Baum–Welch-based, K-means-based, and transfer learning differential evolution-based hidden Markov multi-step attack detection methods.

Mutation-based fuzzing is currently one of the most effective techniques to discover software vulnerabilities. It relies on mutation strategies to generate interesting seeds. As a state-of-the-art mutation-based fuzzer, AFL follows a mutation strategy with high randomization, which uses randomly selected mutation operators to mutate seeds at random offsets. Its strategy may ignore some efficient mutation operators and mutation positions. Therefore, in this paper, we propose a solution named GSA-Fuzz to improve the efficiency of seed mutation strategy with the gravitational search algorithm (GSA). GSA-Fuzz uses GSA to learn the optimal selection probability distributions of operators and mutation positions and designs a position-sensitive strategy to guide seed mutation with learned distributions. Besides, GSA-Fuzz also provides a flip mode to calculate the efficiencies of the deterministic stage and indeterministic stage and implements switching between the two stages to further improve the efficiency of seed mutation. We compare GSA-Fuzz with the state-of-the-art fuzzers AFL, MOPT-AFL, and EcoFuzz on 10 open-source programs. GSA-Fuzz finds 145 more paths than AFL, 66 more paths than EcoFuzz, and 43 more paths than MOPT-AFL. In addition, GSA-Fuzz also outperforms other fuzzers in bug detection and line coverage.

Fuzzing has become an important method for finding vulnerabilities in software. For fuzzing programs expecting structural inputs, syntactic- and semantic-aware fuzzing approaches have been particularly proposed. However, they still cannot fuzz in-memory data stores sufficiently, since some code paths are only executed when the required data are available. In this article, we propose a data-aware fuzzing method, DAFuzz, which is designed by considering the data used during fuzzing. Specifically, to ensure different data-sensitive code paths are exercised, DAFuzz first loads different kinds of data into the stores before feeding fuzzing inputs. Then, when generating inputs, DAFuzz ensures the generated inputs are not only syntactically and semantically valid but also use the data correctly. We implement a prototype of DAFuzz based on Superion and use it to fuzz Redis and Memcached. Experiments show that DAFuzz covers 13~95% more edges than AFL, Superion, AFL++, and AFLNet, and discovers vulnerabilities over 2.7× faster. In total, we discovered four new vulnerabilities in Redis and Memcached. All the vulnerabilities were reported to developers and have been acknowledged and fixed.