We present a rootkit prevention system, namely DARK that tracks suspicious Linux loadable kernel modules (LKM) at a granular level by using on-demand emulation, a technique that dynamically switches a running system between virtualized and emulated execution. Combining the strengths of emulation and virtualization, DARK is able to thoroughly capture the activities of the target module in a guest operating system (OS), while maintaining reasonable run-time performance. To address integrity-violation and confidentiality-violation rootkits, we create a group of security policies that can detect all available Linux rootkits. It is shown that normal guest OS performance is unaffected. The performance is only decreased when rootkits attempt to run, while most rootkits are detected at installation. Next, we present a sandbox-based malware analysis system called Rkprofiler that dynamically monitors and analyzes the behavior of Windows kernel malware. Kernel malware samples run inside a virtual machine (VM) that is supported and managed by a PC emulator. Rkprofiler provides several capabilities that other malware analysis systems do not have. First, it can detect the execution of malicious kernel code regardless of how the monitored kernel malware is loaded into the kernel and whether it is packed or not. Second, it captures all function calls made by the kernel malware and constructs call graphs from the trace files. Third, a technique called aggressive memory tagging (AMT) is proposed to track the dynamic data objects that the kernel malware visits. Last, Rkprofiler records and reports the hardware access events of kernel malware (e.g., MSR register reads and writes). Our evaluation results show that Rkprofiler can quickly expose the security-sensitive activities of kernel malware and thus reduces the effort exerted in conducting tedious manual malware analysis.

Today, host-based malware detection approaches such as antivirus programs are severely lagging in terms of defense against malware. Two important aspects that the overall effectiveness of malware detection depend on are the success of extracting information from malware using malware analysis to generate signatures, and then the success of utilizing these signatures on target hosts with appropriate system monitoring techniques. Today's malware employ a vast array of anti-analysis and anti-monitoring techniques to deter analysis and to neutralize antivirus programs, reducing the overall success of malware detection. In this dissertation, we present a set of practical approaches of robust and efficient malware analysis and system monitoring that can help make malware detection on hosts become more effective. First, we present a framework called Eureka, which efficiently deobfuscates single-pass and multi-pass packed binaries and restores obfuscated API calls, providing a basis for extracting comprehensive information from the malware using further static analysis. Second, we present the formal framework of transparent malware analysis and Ether, a dynamic malware analysis environment based on this framework that provides transparent fine-(single instruction) and coarse-(system call) granularity tracing. Third, we introduce an input-based obfuscation technique that hides trigger-based behavior from any input-oblivious analyzer. Fourth, we present an approach that automatically reverse-engineers the emulator and extracts the syntax and semantics of the bytecode language, which helps constructing control-flow graphs of the bytecode program and enables further analysis on the malicious code. Finally, we present Secure In-VM Monitoring, an approach of efficiently monitoring a target host while being robust against unknown malware that may attempt to neutralize security tools.

In the first essay, I examine value created through external acquisition of nascent technology innovation. External acquisition of new technology is a growing trend in the innovation process, particularly in high technology industries, as firms complement internal efforts with aggressive acquisition programs. Yet, despite its importance, there is little empirical research on the timing of acquisition decisions in high technology environments. I examine the impact of target age on value created for the buyer. Applying an event study methodology to technology acquisitions in the telecommunications industry from 1995 to 2001, empirical evidence supports acquiring early in the face of uncertainty. The equity markets reward the acquisition of younger companies. In sharp contrast to the first essay, the second essay examines the diffusion of negative innovations. While destruction can be creative, certainly not all destruction is creative. Some is just destruction. I examine two fundamentally different paths to information security compromise an opportunistic path and a deliberate path. Through a grounded approach using interviews, observations, and secondary data, I advance a model of the information security compromise process. Using one year of alert data from intrusion detection devices, empirical analysis provides evidence that these paths follow two distinct, but interrelated diffusion patterns. Although distinct, I find empirical evidence that these paths both converge and escalate. Beyond the specific findings in the Internet security context, the study leads to a richer understanding of the diffusion of negative technological innovation. In the third essay, I build on the second essay by examining the effectiveness of reward-based mechanisms in restricting the diffusion of negative innovations. Concerns have been raised that reward-based private infomediaries introduce information leakage which decreases social welfare. Using two years of alert data, I find evidence of their effectiveness despite any leakage which may be occurring. While reward-based disclosures are just as likely to be exploited as non-reward-baed disclosures, exploits from reward-based disclosures are less likely to occur in the first week after disclosure. Further the overall volume of alerts is reduced. This research helps determine the effectiveness of reward mechanisms and provides guidance for security policy makers.

Thesis (M. S.)--Management, Georgia Institute of Technology, 2008. / Committee Chair: Sabyasachi Mitra; Committee Member: Frank Rothaermel; Committee Member: Sandra Slaughter; Committee Member: Sridhar Narasimhan; Committee Member: Vivek Ghosal.

M.Sc. / Anyone who uses a computer for work or recreational purposes has come across one or all of the following problems directly or indirectly (knowingly or not): viruses, worms, trojans, rootkits and botnets. This is especially the case if the computer is connected to the Internet. Looking at the statistics in [1] we can see that although malware detection techniques are detecting and preventing malware, they do not guarantee a 100% detection and or prevention of malware. Furthermore the statistics in [2] show that malware infection rates are increasing around the world at an alarming rate. The statistics also show that there are a high number of new malware samples being discovered every month and that 31% of malware attacks resulted in data loss [3], with 10% of companies reporting the loss of sensitive business data [4][5]. The reason for not being able to achieve a 100% detection and / or prevention of malware is because malware authors make use of sophisticated techniques such as code obfuscation in order to prevent malware from being detected. This has resulted in the emergence of malware known as polymorphic and metamorphic malware. The aforementioned malware poses serious challenges for anti-malware software specifically signature based techniques. However a more serious threat that needs to be addressed is that of rootkits. Rootkits can execute at the same privilege level as the Operating System (OS) itself. At this level the rootkit can manipulate the OS such that it can distribute other malware, hide existing malware, steal information, hide itself, disable anti-malware software etc all without the knowledge of the user. It is clear from the statistics that anti-malware products are not working because infection rates continue to rise and companies and end users continue to fall victims of these attacks. Therefore this dissertation will address the problem that current anti-malware techniques are not working. The main objective of this dissertation is to create a framework called ATE (Anti-malware Technique Evaluator) that can be used to critically evaluate current commercial anti-malware products. The framework will achieve this by identifying the current vulnerabilities that exist in commercial anti-malware products and the operating system. The prior will be achieved by making use of two rootkits, the Evader rootkit and the Sabotager rootkit, which were specifically developed to support the anti-malware product evaluation. Finally an anti-malware architecture we called External Malware Scanner (EMS), will be proposed to address the identified vulnerabilities.

As the number of computing devices connected to the Internet increases and the Internet itself becomes more pervasive, so does the opportunity for criminals to use these devices in cybercrimes. Supporting the increase in cybercrime is the growth and maturity of the digital underground economy with strong links to its more visible and physical counterpart. The digital underground economy provides software and related services to equip the entrepreneurial cybercriminal with the appropriate skills and required tools. Financial malware, particularly the capability for injection of code into web browsers, has become one of the more profitable cybercrime tool sets due to its versatility and adaptability when targeting clients of institutions with an online presence, both in and outside of the financial industry. There are numerous families of financial malware available for use, with perhaps the most prevalent being Zeus and SpyEye. Criminals create (or purchase) and grow botnets of computing devices infected with financial malware that has been configured to attack clients of certain websites. In the research data set there are 483 configuration files containing approximately 40 000 webinjects that were captured from various financial malware botnets between October 2010 and June 2012. They were processed and analysed to determine the methods used by criminals to defraud either the user of the computing device, or the institution of which the user is a client. The configuration files contain the injection code that is executed in the web browser to create a surrogate interface, which is then used by the criminal to interact with the user and institution in order to commit fraud. Demographics on the captured data set are presented and case studies are documented based on the various methods used to defraud and bypass financial security controls across multiple industries. The case studies cover techniques used in social engineering, bypassing security controls and automated transfers.

M. Tech. (Department of Information Technology, Faculty of Applied and Computer Sciences), Vaal University of Technology / Malware is today one of the biggest security threats to the Internet. Malware is any malicious software with the intent to perform malevolent activities on a targeted system. Viruses, worms, trojans, backdoors and adware are but a few examples that fall under the umbrella of malware. The purpose of this research is to investigate techniques that are used in order to effectively perform Malware analysis and detection on enterprise systems to reduce the damage of malware attacks on the operation of organizations. Malware analysis experiments were carried out using the two techniques of malware analysis, which are Dynamic and Static analysis, on two different malware samples. Portable executable and Microsoft word document files were the two samples that were analysed in an isolated sandbox lab environment. Static analysis is the process of examining and extracting information from malware code without executing the malware, while Dynamic analysis is the process of executing malware in order to observe and record its behaviour in a controlled environment. The results from the experiments disclosed the behaviour, encryption techniques, and other techniques employed by the malware samples. These malware analysis experiments were carried out in an isolated lab environment that was built for the purpose of this research. The results showed that Dynamic analysis is more effective than Static analysis. The study proposes the use of both techniques for comprehensive malware analysis and detection.

Security vulnerabilities have a large real-world impact, from ransomware attacks costing billions of dollars every year to sensitive data breaches in government, military and industry. Fuzzing is a popular technique to discover these vulnerabilities in an automated fashion. Industries have poured tons of resources into building large-scale fuzzing factories (e.g., Google’s ClusterFuzz and Microsoft’s OneFuzz) to test their products and make their product more secure. Despite the wide application of fuzzing in industry, there remain many issues constraining its performance. One fundamental limitation is the rule-based design in fuzzing. Rule-based fuzzers heavily rely on a set of static rules or heuristics. These fixed rules are summarized from human experience, hence failing to generalize on a diverse set of programs. In this dissertation, we present an adaptive and effective fuzzing framework in data-driven approach. A data-driven fuzzer makes decisions based on the analysis and reasoning of data rather than the static rules. Hence it is more adaptive, effective, and flexible than a typical rule-based fuzzer. More interestingly, the data-driven approach can bridge the connection from fuzzing to various data-centric domains (e.g., machine learning, optimizations and social network), enabling sophisticated designs in the fuzzing framework. A general fuzzing framework consists of two major components: seed scheduling and seed mutation. The seed scheduling module selects a seed from a seed corpus that includes multiple testcases. Then seed mutation module applies perturbation on the selected seed to generate a new testcase. First, we present Neuzz, the first machine learning (ML) based general-purpose fuzzer that adopts ML to seed mutation and greatly improves fuzzing performance. Then we present MTFuzz, a follow-up work of Neuzz by including diverse data into ML to generate effective seed mutations. In the end, we present K-Scheduler, a fuzzer-agnostic seed scheduling algorithm in data-driven approach. K-Scheduler leverages the graph data (i.e., inter-procedural control flow graph) and dynamic coverage data (i.e., code coverage bitmap) to construct a dynamic graph and schedule seeds by the graph centrality scores on that graph. It can significantly improve the fuzzing performance than the-state-of-art seed schedulers on various fuzzers widely-used in the industry.

Worldwide computer systems continue to execute malicious software that degrades the systemsâ performance and consumes network capacity by generating high volumes of unwanted traffic. Network-based detectors can effectively identify machines participating in the ongoing attacks by monitoring the traffic to and from the systems. But, network detection alone is not enough; it does not improve the operation of the Internet or the health of other machines connected to the network. We must identify malicious code running on infected systems, participating in global attack networks. This dissertation describes a robust and secure approach that identifies malware present on infected systems based on its undesirable use of network. Our approach, using virtualization, attributes malicious traffic to host-level processes responsible for the traffic. The attribution identifies on-host processes, but malware instances often exhibit parasitic behaviors to subvert the execution of benign processes. We then augment the attribution software with a host-level monitor that detects parasitic behaviors occurring at the user- and kernel-level. User-level parasitic attack detection happens via the system-call interface because it is a non-bypassable interface for user-level processes. Due to the unavailability of one such interface inside the kernel for drivers, we create a new driver monitoring interface inside the kernel to detect parasitic attacks occurring through this interface. Our attribution software relies on a guest kernelâ s data to identify on-host processes. To allow secure attribution, we prevent illegal modifications of critical kernel data from kernel-level malware. Together, our contributions produce a unified research outcome --an improved malicious code identification system for user- and kernel-level malware.

The Domain Name System (DNS) is a critical component of the Internet. DNS provides the ability to map human-readable and memorable domain names to machine-level IP addresses and other records. These mappings lie at the heart of the Internet's success and are essential for the majority of core Internet applications and protocols. The critical nature of DNS means that it is often the target of abuse. Cyber-criminals rely heavily upon the reliability and scalability of the DNS protocol to serve as an agile platform for their illicit operations. For example, modern malware and Internet fraud techniques rely upon DNS to locate their remote command-and-control (C&C) servers through which new commands from the attacker are issued, serve as exfiltration points for information stolen from the victims' computers, and to manage subsequent updates to their malicious toolset. The research described in this thesis scientifically addresses problems in the area of DNS-based detection of illicit operations. In detail, this research studies new methods to quantify and track dynamically changing reputations for DNS based on passive network measurements. The research also investigates methods for the creation of early warning systems for DNS. These early warning systems enables the research community to identify emerging threats (e.g., new botnets and malware infections) across the DNS hierarchy in a timelier manner.