Hervé Debar

Aggregation and Correlation of Intrusion-Detection Alerts

This paper describes an aggregation and correlation algorithm used in the design and implementation of an intrusion-detection console built on top of the Tivoli Enterprise Console (TEC). The aggregation and correlation algorithm aims at acquiring intrusion-detection alerts and relating them together to expose a more condensed view of the security issues raised by intrusion-detection systems.

M2D2: a formal data model for IDS alert correlation

At present, alert correlation techniques do not make full use of the information that is available. We propose a data model for IDS alert correlation called M2D2. It supplies four information types: information related to the characteristics of the monitored information system, information about the vulnerabilities, information about the security tools used for the monitoring, and information about the events observed. M2D2 is formally defined. As far as we know, no other formal model includes the vulnerability and alert parts of M2D2. Three examples of correlations are given. They are rigorously specified using the formal definition of M2D2. As opposed to already published correlation methods, these examples use more than the events generated by security tools; they make use of many concepts formalized in M2D2.

Correlation of Intrusion Symptoms: An Application of Chronicles

In this paper, we propose a multi-alarm misuse correlation component based on the chronicles formalism. Chronicles provide a high level declarative language and a recognition system that is used in other areas where dynamic systems are monitored. This formalism allows us to reduce the number of alarms shipped to the operator and enhances the quality of the diagnosis provided.

Behavioral detection of malware: from a survey towards an established taxonomy

Behavioral detection differs from appearance detection in that it identifies the actions performed by the malware rather than syntactic markers. Identifying these malicious actions and interpreting their final purpose is a complex reasoning process. This paper draws up a survey of the different reasoning techniques deployed among the behavioral detectors. These detectors have been classified according to a new taxonomy introduced inside the paper. Strongly inspired from the domain of program testing, this taxonomy divides the behavioral detectors into two main families: simulation-based and formal detectors. Inside these families, ramifications are then derived according to the data collection mechanisms the data interpretation, the adopted model and its generation, and the decision support.

A service dependency model for cost-sensitive intrusion response

Recent advances in intrusion detection and prevention have brought promising solutions to enhance IT security. Despite these efforts, the battle with cyber attackers has reached a deadlock. While attackers always try to unveil new vulnerabilities, security experts are bounded to keep their softwares compliant with the latest updates. Intrusion response systems are thus relegated to a second rank because no one trusts them to modify system configuration during runtime. Current response cost evaluation techniques do not cover all impact aspects, favoring availability over confidentiality and integrity. They do not profit from the findings in intrusion prevention which led to powerful models including vulnerability graphs, exploit graphs, etc. This paper bridges the gap between these models and service dependency models that are used for response evaluation. It proposes a new service dependency representation that enables intrusion and response impact evaluation. The outcome is a service dependency model and a complete methodology to use this model in order to evaluate intrusion and response costs. The latter covers response collateral damages and positive response effects as they reduce intrusion costs.

A logic-based model to support alert correlation in intrusion detection

Managing and supervising security in large networks has become a challenging task, as new threats and flaws are being discovered on a daily basis. This requires an in depth and up-to-date knowledge of the context in which security-related events occur. Several tools have been proposed to support security operators in this task, each of which focuses on some specific aspects of the monitoring. Many alarm fusion and correlation approaches have also been investigated. However, most of these approaches suffer from two major drawbacks. First, they only take advantage of the information found in alerts, which is not sufficient to achieve the goals of alert correlation, that is to say to reduce the overall amount of alerts, while enhancing their semantics. Second, these techniques have been designed on an ad hoc basis and lack a shared data model that would allow them to reason about events in a cooperative way. In this paper, we propose a federative data model for security systems to query and assert knowledge about security incidents and the context in which they occur. This model constitutes a consistent and formal ground to represent information that is required to reason about complementary evidences, in order to confirm or invalidate alerts raised by intrusion detection systems.

Time series modeling for IDS alert management

Intrusion detection systems create large amounts of alerts. Significant part of these alerts can be seen as background noise of an operational information system, and its quantity typically overwhelms the user. In this paper we have three points to make. First, we present our findings regarding the causes of this noise. Second, we provide some reasoning why one would like to keep an eye on the noise despite the large number of alerts. Finally, one approach for monitoring the noise with reasonable user load is proposed. The approach is based on modeling regularities in alert flows with classical time series methods. We present experimentations and results obtained using real world data.

Processing intrusion detection alert aggregates with time series modeling

The main use of intrusion detection systems (IDS) is to detect attacks against information systems and networks. Normal use of the network and its functioning can also be monitored with an IDS. It can be used to control, for example, the use of management and signaling protocols, or the network traffic related to some less critical aspects of system policies. These complementary usages can generate large numbers of alerts, but still, in operational environment, the collection of such data may be mandated by the security policy. Processing this type of alerts presents a different problem than correlating alerts directly related to attacks or filtering incorrectly issued alerts. We aggregate individual alerts to alert flows, and then process the flows instead of individual alerts for two reasons. First, this is necessary to cope with the large quantity of alerts - a common problem among all alert correlation approaches. Second, individual alerts relevancy is often indeterminable, but irrelevant alerts and interesting phenomena can be identified at the flow level. This is the particularity of the alerts created by the complementary uses of IDSes. Flows consisting of alerts related to normal system behavior can contain strong regularities. We propose to model these regularities using non-stationary autoregressive models. Once modeled, the regularities can be filtered out to relieve the security operator from manual analysis of true, but low impact alerts. We present experimental results using these models to process voluminous alert flows from an operational network.

An extended RBAC profile of XACML

Nowadays many organizations use security policies to control access to sensitive resources. Moreover, exchanging or sharing services and resources is essential for these organizations to achieve their business objectives. Since the eXtensible Access Control Markup Language (XACML) was standardized by the OASIS community, it has been widely deployed, making it easier to interoperate with other applications using the same standard language. The OASIS has defined an RBAC profile of XACML that illustrates how organizations that would like to use the RBAC model can express their access control policy within this standard language. This work analyzes the RBAC profile of XACML, showing its limitations to respond to all the requirements for access control. We then suggest adding some functionalities within an extended RBAC profile of XACML. This new profile is expected to respond to more advanced access control requirements such as user-user delegation, access elements abstractions and contextual applicability of the policies.

Fixed vs. Variable-Length Patterns for Detecting Suspicious Process Behavior

This paper addresses the problem of creating patterns that can be used to model the normal behavior of a given process. These models can be used for intrusion detection purposes. In a previous work, we presented a novel method to generate input data sets that enable us to observe the normal behavior of a process in a secure environment. Using this method, we propose various techniques to generate either fixed-length or variable-length patterns. We show the advantages and drawbacks of each technique, based on the results of the experiments we have run on our testbed.