Xi Xiong

SHELF: Preserving Business Continuity and Availability in an Intrusion Recovery System

Recovering from intrusions for a compromised computer system is a challenging job, especially for systems that run continuous services. Current intrusion recovery techniques often do not preserve the accumulated useful state of running applications and have very limited system availability when performing recovery routines. In this paper, we propose SHELF, an on-the-fly intrusion recovery prototype system that provides a comprehensive solution to preserve business continuity, availability and recovery accuracy. SHELF preserves accumulated clean states for infected applications and files so that they can continue with the most recent pre-infection states after recovery. Moreover, SHELF leverages OS-aware taint tracking techniques to swiftly determine the sources of intrusion and assess system-wide damages caused by the intrusion. SHELF uses quarantine methods to prevent infection propagation so that uninfected and recovered objects can provide availability during the recovery phase. We integrate SHELF prototype in a virtualization environment to achieve user transparency and protection. Our evaluation shows that SHELF can perform accurate recovery on-the-fly effectively with an acceptable performance overhead.

Cross-Layer Damage Assessment for Cyber Situational Awareness

Damage assessment plays a very important role in securing enterprise networks and systems. Gaining good awareness about the effects and impact of cyber attack actions would enable security officers to make the right cyber defense decisions and take the right cyber defense actions. A good number of damage assessment techniques have been proposed in the literature, but they typically focus on a single abstraction level (of the software system in concern). As a result, existing damage assessment techniques and tools are still very limited in satisfying the needs of comprehensive damage assessment which should not result in any “blind spots”.

Defeating buffer overflow attacks via virtualization

Display Omitted We propose an on-the-fly buffer overflow prevention mechanism, which can protect the target program without restarting it.We make use of the virtualization technology to transparently defend against buffer overflow attacks.Our system does not need any changes to the existing programs and OS, and it can be easily deployed in the VM environment. Buffer overflow defenses have been comprehensively studied for many years. Different from previous solutions, we propose PHUKO, an on-the-fly buffer overflow prevention system which leverages virtualization technology. PHUKO offers the protected program a fully transparent environment and an easy deployment without the need to restart the program. Generally, the working process of PHUKO can be divided into two stages. First, we utilize static binary analysis to identify the instructions offline which are the entries of vulnerable functions. Second, by combining virtual machine introspection and online patching, PHUKO instruments the protected running program on-the-fly with memory safety enforcement. The experiments show that our system can defend against realistic buffer overflow attacks effectively with a moderate performance overhead.

SILVER: Fine-Grained and Transparent Protection Domain Primitives in Commodity OS Kernel

Untrusted kernel extensions remain one of the major threats to the security of commodity OS kernels. Current containment approaches still have limitations in terms of security, granularity and flexibility, primarily due to the absence of secure resource management and communication methods. This paper presents SILVER, a framework that offers transparent protection domain primitives to achieve fine-grained access control and secure communication between OS kernel and extensions. SILVER keeps track of security properties e.g., owner principal and integrity level of data objects in kernel space with a novel security-aware memory management scheme, which enables fine-grained access control in an effective manner. Moreover, SILVER introduces secure primitives for data communication between protection domains based on a unified integrity model. SILVERs protection domain primitives provide great flexibility by allowing developers to explicitly define security properties of individual program data, as well as control privilege delegation, data transfer and service exportation. We have implemented a prototype of SILVER in Linux. The evaluation results reveal that SILVER is effective against various kinds of kernel threats with a reasonable performance and resource overhead.

Policy-centric protection of OS kernel from vulnerable loadable kernel modules

Due to lack of the protecting mechanism in the kernel space, the loadable kernel modules (LKM) may be exploited and thus seriously affecting the OS kernels security via utilizing the implicit or explicit vulnerabilities. Although lots of systems have been developed to address the above problem, there still remain some challenges. a) How to automatically generate a security policy before the kernel module is enforced? b) How to properly mediate the interactions between the kernel module and OS kernel to ensure the policy consistence without modifications (or least changes) on the existing OS, hardware, and kernel module structure? In this paper, we present LKMG, a policy-centric system which can protect commodity OS kernel from vulnerable loadable kernel modules. More powerful than previous systems, LKMG is able to generate a security policy form the kernel module, and then enforce the policy during the kernel modules execution. Generally, the working process of LKMG can be divided into two stages. First, we utilize static analysis to extract the kernel code and data access patterns from a kernel modules source code, and then combine these patterns with the related memory address information to generate a security policy. Second, by leveraging hardware-based virtualization technology, LKMG isolates the kernel module from the rest of the kernel, and then enforces the kernel modules execution to obey the derived policy. The experiment show that our system can defend against various loadable kernel module exploitations effectively with moderate performance overhead.

Integrating offline analysis and online protection to defeat buffer overflow attacks

Nowadays Buffer overflow attacks are still recognized as one of the most severe threats in software security. Previous solutions suffer from limitations in that: 1) Some methods based on compiler extensions have limited practicality because they need to access source code; 2) Other methods that need to modify some aspects of the operating system or hardware require much deployment effort; 3) Almost all methods are unable to deploy a runtime protection for programs that cannot afford to restart. In this paper, we propose PHUKO, an on-the-fly buffer overflow prevention system which leverages virtualization technology. PHUKO offers the protected program a fully transparent environment and an easy deployment without the need to restart the program. The experiments show that our system can defend against realistic buffer overflow attacks effectively with moderate performance overhead.

Using purpose capturing signatures to defeat computer virus mutating

Nowadays computer viruses become more and more difficult to be identified. Modern computer viruses use various mutation techniques such as polymorphism and metamorphism to evade detection. Previous researches in mutated computer virus detection have limitations in that: 1) most of them cannot handle advanced mutation techniques; 2) the methods based on source code analysis are less practical. 3) some methods are unable to detect computer viruses immediately. In this paper, we present a new dynamic approach to detect and analyze computer viruses based on Virtual Machine technology. We show that 1) how to generate Purpose Capturing Signatures based on the information of runtime values (execution value sequence, EVS) and control flows (execution control sequence, ECS); 2) how to detect and analyze computer viruses using the purpose-capturing signatures. To our best knowledge, it is the first method to perform computer virus detection and analysis using the EVS and ECS. Our experimental evaluation demonstrates that this approach is able to use one signature to detect all mutations of the corresponding virus efficiently.