Ebook: A Systems Approach to Cyber Security

Welcome to SG-CRC 2017, the second Singapore Cyber Security R&D Conference and welcome to Singapore! This year's theme focuses on presenting how to approach Cyber-Security in complex computing systems. Thus the focus of this year's theme is on software and systems security. In order to preserve information security of systems it is necessary to consider a wide variety of techniques to avoid cyber-attacks, or to minimize potential damage caused by a successful attacker. SG-CRC 2017 focuses on techniques and methodologies oriented to construct resilient systems against cyber-attacks that helps to construct safe execution environments, improving security of both hardware and software by means of using mathematical tools and engineering approaches for designing, verifying, and monitoring cyber-physical systems.

This year's conference contains a strong technical program. In addition to three keynotes addresses by Ruby Lee, Luke Ong and Sjouke Mauw, it also includes a selection of rigorously refereed papers presented in the regular paper sessions and short paper sessions. The Program Committee has received 21 submissions internationally and each paper was reviewed by at least 3 referees. We choose 10 papers as the result of intensive discussions held among the PC members.

This year, a two-day Cybersecurity Camp is co-located with the conference, which is organized by Singapore Cybersecurity Consortium. We would like to thank Dawn Song for sharing her expertise in cyber-security with students and attendees in the cyber-camp via lectures and hackathon. This year conference has also seen the formal launch of the National Cyber-security Lab, a national infra-structure housed at the National University of Singapore.

We would like to thank National University of Singapore and National Research Foundation as co-organizers of the conference. Special thanks go to many individuals who have contributed to the success of this conference. We thank the authors for sharing their ideas with us, the reviewers for providing valuable feedback, and all the PC members for taking time from their busy schedules to support this conference.

Thank you so much for attending SG-CRC 2017. We hope that you enjoy the program and have a great stay in Singapore!

Abhik Roychoudhury and Yang Liu (Program Co-Chairs)

Hwee Kwang Lim (General Chair)

Recent years have seen a pervasive usage of mobile-based instant messaging apps, which are popularly known as chat apps. On users' mobile devices, chat logs are usually stored encrypted. This paper is concerned with discovering the decryption key of chat-log database files as they are used by popular chat apps like WhatsApp and WeChat. We propose a systematic and generalized information-flow based approach to recovering the decryption key by taking advantage of both static and dynamic analyses. We show that, despite the employed code obfuscation techniques, we can perform the key discovery process on relevant code portions. Furthermore, to the best of our knowledge, we are the first to detail the employed string de-obfuscation, encrypted database file structure, and decryption-key formulation of the latest WhatsApp with crypt12 database. We also demonstrate how our key-extraction techniques can decrypt encrypted WhatsApp and WeChat database files that originate from a target device. Additionally, we show how we can construct a version of WhatsApp or WeChat that simulates the key generation processes of a remote target device, and recover the keys. Lastly, we analyze why our technique can work on widely-popular chat apps, and mention measures that can be adopted by chat-app developers to better protect the privacy of billions of their users.

The high security requirements of cyber-physical systems and the critical tasks they carry out make it necessary to guarantee the absence of any vulnerability to security attacks and that they have no unexpected behaviour. The size and complexity of the underlying hardware in cyber-physical systems are increasing and so is the risk of failures and vulnerability to security threats. Checking for errors and security holes in the early phases of hardware production is generally considered an effective approach. To ease the process of designing and testing in the early stage, hardware description languages such as Verilog are used to design hardware. Hardware designs in such description languages, however, do not correspond to mathematical models. Hence we cannot reason about hardware designs and formally verify their correctness and security related properties. In this paper, we develop a formal model of Verilog in the theorem prover Isabelle/HOL. Our model covers most constructs used in hardware designs. With our model, one can analyse Verilog designs and execute them for simulation. More importantly, our model enables formal reasoning about interesting properties for Verilog designs. To complete our tool chain, we build a translator which automatically translates an existing Verilog design to equivalent design in our formal model.

As electric sector stakeholders make the decision to upgrade traditional power grid architectures by incorporating smart grid technologies and new intelligent components, the benefits of added connectivity must be weighed against the risk of increased exposure to cyber attacks. Therefore, decision makers must ask: how smart is smart enough? We present a probabilistic risk analysis approach to this problem. Central to this approach is a new network security model based on a reformulation of the classic “multi-armed bandits” problem, where instead of projects with uncertain probabilities of success, a network defender faces network nodes that can be attacked at uncertain Poisson-distributed rates. Probing these nodes provides additional information about their vulnerability, but at a cost. Using this new technique, which by similarity we call “multi-node bandits”, we compute the net marginal benefits of increased connectivity. We illustrate this model by the quantification of the overall cyber risk to the physical and informational networks of a smart grid in order to identify the optimal degree of “smartness” and the best risk management strategy.

Web forums are a frequent way of sharing useful information among people. They are becoming the main source of up-to-date information and marketplaces pertaining to different domains, including criminal content and zero-day security exploits. Analyzing the web forums of the existing discussion threads is an alternative method to understand the exploits and fraud modalities a law breaker will most likely make use and how to defend against them. However, in many cases, it is hard to capture all the relevant context of the forums which is needed for classification. In this paper, we introduce a data-driven technique to mine the web forums and provide policy recommendations to the defender. A neural network (NN) is used to learn the set of features for forum classification. Furthermore, we present the evaluation and results from employing our method, with various system configurations, on real-world datasets collected form the web.

A smooth operation of the Platform Screen Door (PSD) system is critical to the Metro system, and any disturbance to it may disrupt the train's normal operation. The paper presents the security analysis of a Metro Supervisory Control and Data Acquisition (SCADA) system, specifically the cyber security vulnerabilities in its PSD system. The PSD system includes control subsystem and signaling subsystem, and its operation can be controlled from the moving train and the station as well. The security features of communication protocols that are employed in the SCADA system and PSD system operation mechanisms are discussed in this paper. The weak security features render the PSD vulnerable to cyber attacks. Countermeasures, from both technical and human aspects, to protect the PSD system are studied. An experiment is conducted on a testbed of simulating Metro supervisory control system to test the system vulnerabilities. The results demonstrate that the PSD control system could be compromised by an attacker who gains physical access to the control network and launches forged message or replay message attacks. A firewall cyber security countermeasure is evaluated to show that it can prevent some of the attacks but has limitations due to its rule-based mechanism. Thus, it is necessary to mind the gap for the security of metro PSD system.

Messaging systems operating within the public cloud are gaining popularity. To protect message confidentiality from the public cloud including the public messaging servers, we propose to encrypt messages in messaging systems using Attribute-Based Encryption (ABE). ABE is an one-to-many public key encryption system in which data are encrypted with access policies and only users with attributes that satisfy the access policies can decrypt the ciphertexts, and hence is considered as a promising solution for realizing expressive and fine-grained access control of encrypted data in public servers. Our proposed system, called Attribute-Based Secure Messaging System with Outsourced Decryption (ABSM-OD), has three key features: enabling expressive and fine-grained access control of encrypted messages by users, supporting outsourced decryption to the cloud while without compromising confidentiality of decrypted messages, and allowing server-aided revocation to provide effective and instant user revocations.

In this paper, we show efficient elliptic curve cryptography implementations over advanced ARMv8 processor. We improve the previous binary field multiplication over the processor with finely aligned multiplication and incomplete reduction techniques by taking advantages of advanced 64-bit polynomial multiplication (PMULL). This approach shows performance enhancements by a factor of 1.34 times than previous implementation of binary field multiplication. For the point addition and doubling, the special types of multiplication, squaring and addition operations are combined together and optimized, where one reduction operation is optimized in each case. The scalar multiplication is implemented in constant-time window method, which is secure against timing attacks. Finally the proposed implementations achieved 759,630/331,944 clock cycles for random/fixed scalar multiplications for B-571 curve over ARMv8, respectively.

Researchers from academic institutions and the corporate sector rely heavily on scholarly digital libraries for accessing journal articles and conference proceedings. Primarily downloaded in the form of PDF files, there is a risk that these documents may be compromised by attackers. PDF files have many capabilities that have been widely used for malicious operations. Attackers increasingly take advantage of innocent users who open PDF files with little or no concern, mistakenly considering these files safe and relatively non-threatening. Researchers also consider scholarly digital libraries reliable and home to a trusted corpus of papers and untainted by malicious files. For these reasons, scholarly digital libraries are an attractive target for cyber-attacks launched via PDF files. In this study, we present several vulnerabilities and practical distribution attack approaches tailored for scholarly digital libraries. To support our claim regarding the attractiveness of scholarly digital libraries as an attack platform, we evaluated more than two million scholarly papers in the CiteSeerX library that were collected over 8 years and found it to be contaminated with a surprisingly large number (0.3%–2%) of malicious scholarly PDF documents, the origin of which is 46 different countries spread worldwide. More than 55% of the malicious papers in CiteSeerX were crawled from IP's belonging to USA universities, followed by those belonging to Europe (33.6%). We show how existing scholarly digital libraries can be easily leveraged as a distribution platform both for a targeted attack and in a worldwide manner. On average, a certain malicious paper caused high impact damage as it was downloaded 167 times in 5 years by researchers from different countries worldwide. In general, the USA and Asia downloaded the most malicious scholarly papers, 40.15% and 27.9%, respectively. The top malicious scholarly document downloaded is a malicious version of a popular paper in the computer forensics domain, with 2213 downloads in a worldwide coverage of 108 different countries. Finally, we suggest several concrete solutions for mitigating such attacks, including simple deterministic solutions and also advanced machine learning-based frameworks.

File carving is the process which aims to recover files from storage media without the file system meta-data. The ability to perform such recovery is particularly important in this digital era when it involves forensic investigation. Due to the inevitable occurrence of file fragmentation in storage system, fragment classification is an important step in the file recovery process. Following the increase of storage capacity and usage of mobile phones, large amount of personal data tends to be stored on such devices, which is of great interest for forensic analysis during investigations. In this paper, we present an approach in classifying the most commonly found fragment types on mobile phones, which include JPG, MP3, MP4, MOV and SQLite. Departing from the conventional approaches that utilize analysis derived from unigram statistics, we employ bigram statistics in our approach in order to capture the frequency of local byte order which retains meaningful and exploitable pattern in the fragments. While being able to capture more information, the bigram statistics also contain a large amount of redundant data which greatly increases the computational workload. Therefore, we perform dimensionality reduction through Principal Component Analysis (PCA) in order to extract only the most significant dimensions for classification purpose of the targeted file types. Using the resulting features, an average classification accuracy of 96.19% is achieved, comparing to 88.40% while using the unigram statistics alone through Support Vector Machine (SVM).

In recent years, hardware obfuscation is one of the prominent anti-tamper solutions that are highly used against various hardware security threats such as piracy, cloning, reverse engineering, chip overbuilding, and hardware Trojans. Logic obfuscation is implemented either in the design description (for soft/firm/hard codes) or structure (for the chip) of electronic hardware to intentionally conceal its functionality. Most of the obfuscation schemes enable the circuit operation in two distinct modes such as obfuscated and normal modes. The mode control, mostly implemented by finite state machine, is performed by the application of a specific sequence of input vectors on initialization, called an ‘initialization key.’ Without the initialization key, it is difficult to comprehend the intended functional behavior of the circuit; hence, circuit tampering or malicious insertion will have a high probability of either becoming functionally benign or easily detectable by conventional logic testing. However, most of the existing obfuscation techniques have used similar obfuscation cell structures throughout the design which in turn leaves a hint to the adversary about circuit obfuscation during reverse engineering. In this paper, we aimed to mitigate this limitation by applying design obfuscation using different obfuscation cells. We performed the hardware obfuscation mechanism for field programmable gate array devices using standard ISCAS'89 benchmark circuits in Actel's ProAsic3 device by Libero SoC v10.1 (free version). We measured the performance overhead using the design parameters such as area, delay, and power.