Anoop Singhal

An Attack Graph-Based Probabilistic Security Metric

To protect critical resources in todays networked environments, it is desirable to quantify the likelihood of potential multi-step attacks that combine multiple vulnerabilities. This now becomes feasible due to a model of causal relationships between vulnerabilities, namely, attack graph. This paper proposes an attack graph-based probabilistic metric for network security and studies its efficient computation. We first define the basic metric and provide an intuitive and meaningful interpretation to the metric. We then study the definition in more complex attack graphs with cycles and extend the definition accordingly. We show that computing the metric directly from its definition is not efficient in many cases and propose heuristics to improve the efficiency of such computation.

Measuring network security using dynamic bayesian network

Given the increasing dependence of our societies on networked information systems, the overall security of these systems should be measured and improved. Existing security metrics have generally focused on measuring individual vulnerabilities without considering their combined effects. Our previous work tackle this issue by exploring the causal relationships between vulnerabilities encoded in an attack graph. However, the evolving nature of vulnerabilities and networks has largely been ignored. In this paper, we propose a Dynamic Bayesian Networks (DBNs)-based model to incorporate temporal factors, such as the availability of exploit codes or patches. Starting from the model, we study two concrete cases to demonstrate the potential applications. This novel model provides a theoretical foundation and a practical framework for continuously measuring network security in a dynamic environment.

Toward measuring network security using attack graphs

In measuring the overall security of a network, a crucial issue is to correctly compose the measure of individual components. Incorrect compositions may lead to misleading results. For example, a network with less vulnerabilities or a more diversified configuration is not necessarily more secure. To obtain correct compositions of individual measures, we need to first understand the interplay between network components. For example, how vulnerabilities can be combined by attackers in advancing an intrusion. Such an understanding becomes possible with recent advances in modeling network security using attack graphs. Based on our experiences with attack graph analysis, we propose an integrated framework for measuring various aspects of network security. We first outline our principles andmethodologies. We then describe concrete examples to buildintuitions. Finally, we present our formal framework. It is our belief that metrics developed based on the proposed framework will lead to novel quantitative approaches to vulnerability analysis, network hardening, and attack response.

k-Zero Day Safety: A Network Security Metric for Measuring the Risk of Unknown Vulnerabilities

By enabling a direct comparison of different security solutions with respect to their relative effectiveness, a network security metric may provide quantifiable evidences to assist security practitioners in securing computer networks. However, research on security metrics has been hindered by difficulties in handling zero-day attacks exploiting unknown vulnerabilities. In fact, the security risk of unknown vulnerabilities has been considered as something unmeasurable due to the less predictable nature of software flaws. This causes a major difficulty to security metrics, because a more secure configuration would be of little value if it were equally susceptible to zero-day attacks. In this paper, we propose a novel security metric, k-zero day safety, to address this issue. Instead of attempting to rank unknown vulnerabilities, our metric counts how many such vulnerabilities would be required for compromising network assets; a larger count implies more security because the likelihood of having more unknown vulnerabilities available, applicable, and exploitable all at the same time will be significantly lower. We formally define the metric, analyze the complexity of computing the metric, devise heuristic algorithms for intractable cases, and finally demonstrate through case studies that applying the metric to existing network security practices may generate actionable knowledge.

k-zero day safety: measuring the security risk of networks against unknown attacks

The security risk of a network against unknown zero day attacks has been considered as something unmeasurable since software flaws are less predictable than hardware faults and the process of finding such flaws and developing exploits seems to be chaotic [10]. In this paper, we propose a novel security metric, k-zero day safety, based on the number of unknown zero day vulnerabilities. That is, the metric simply counts how many unknown vulnerabilities would be required for compromising a network asset, regardless of what vulnerabilities those might be. We formally define the metric based on an abstract model of networks and attacks. We then devise algorithms for computing the metric. Finally, we show the metric can quantify many existing practices in hardening a network.

Aggregating vulnerability metrics in enterprise networks using attack graphs

Quantifying security risk is an important and yet difficult task in enterprise network security management. While metrics exist for individual software vulnerabilities, there is currently no standard way of aggregating such metrics. We present a model that can be used to aggregate vulnerability metrics in an enterprise network, producing quantitative metrics that measure the likelihood breaches can occur within a given network configuration. A clear semantic model for this aggregation is an important first step toward a comprehensive network security metric model. We utilize existing work in attack graphs and apply probabilistic reasoning to produce an aggregation that has clear semantics and sound computation. We ensure that shared dependencies between attack paths have a proportional effect on the final calculation. We correctly reason over cycles, ensuring that privileges are evaluated without any self-referencing effect. We introduce additional modeling artifacts in our probabilistic graphical model to capture and account for hidden correlations among exploit steps. The paper shows that a clear semantic model for aggregation is critical in interpreting the results, calibrating the metric model, and explaining insights gained from empirical evaluation. Our approach has been rigorously evaluated using a number of network models, as well as data from production systems.

Investigating the application of moving target defenses to network security

This paper presents a preliminary design for a moving-target defense (MTD) for computer networks to combat an attackers asymmetric advantage. The MTD system reasons over a set of abstract models that capture the networks configuration and its operational and security goals to select adaptations that maintain the operational integrity of the network. The paper examines both a simple (purely random) MTD system as well as an intelligent MTD system that uses attack indicators to augment adaptation selection. A set of simulation-based experiments show that such an MTD system may in fact be able to reduce an attackers success likelihood. These results are a preliminary step towards understanding and quantifying the impact of MTDs on computer networks.

Interactive analysis of attack graphs using relational queries

Attack graph is important in defending against well-orchestrated network intrusions. However, the current analysis of attack graphs requires an algorithm to be developed and implemented, causing a delay in the availability of analysis. Such a delay is usually unacceptable because the needs for analyzing attack graphs may change rapidly in defending against network intrusions. An administrator may want to revise an analysis upon observing its outcome. Such an interactive analysis, similar to that in decision support systems, is difficult if at all possible with current approaches based on proprietary algorithms. This paper removes the above limitation and enables interactive analysis of attack graphs. We devise a relational model for representing necessary inputs including network configuration and domain knowledge. We generate the attack graph from those inputs as relational views. We then show that typical analyses of the attack graph can be realized as relational queries against the views. Our approach eliminates the needs for developing a proprietary algorithm for each different analysis, because an analysis is now simply a relational query. The interactive analysis of attack graphs is now possible, because relational queries can be dynamically constructed and revised at run time. Moreover, the mature optimization techniques in relational databases can also improve the performance of the analysis.

Security Risk Analysis of Enterprise Networks Using Probabilistic Attack Graphs

Today’s information systems face sophisticated attackers who combine multiple vulnerabilities to penetrate networks with devastating impact. The overall security of an enterprise network cannot be determined by simply counting the number of vulnerabilities. To more accurately assess the security of enterprise systems, one must understand how vulnerabilities can be combined and exploited to stage an attack. Composition of vulnerabilities can be modeled using probabilistic attack graphs, which show all paths of attacks that allow incremental network penetration. Attack likelihoods are propagated through the attack graph, yielding a novel way to measure the security risk of enterprise systems. This metric for risk mitigation analysis is used to maximize the security of enterprise systems. This methodology based on probabilistic attack graphs can be used to evaluate and strengthen the overall security of enterprise networks.

Data warehousing and data mining techniques for intrusion detection systems

This paper describes data mining and data warehousing techniques that can improve the performance and usability of Intrusion Detection Systems (IDS). Current IDS do not provide support for historical data analysis and data summarization. This paper presents techniques to model network traffic and alerts using a multi-dimensional data model and star schemas. This data model was used to perform network security analysis and detect denial of service attacks. Our data model can also be used to handle heterogeneous data sources (e.g. firewall logs, system calls, net-flow data) and enable up to two orders of magnitude faster query response times for analysts as compared to the current state of the art. We have used our techniques to implement a prototype system that is being successfully used at Army Research Labs. Our system has helped the security analyst in detecting intrusions and in historical data analysis for generating reports on trend analysis.