Semir Daskapan

Systems for Detecting Advanced Persistent Threats: A Development Roadmap Using Intelligent Data Analysis

Cyber-attacks against companies and governments are increasing in complexity, persistence and numbers. Common intrusion detection methods lack the ability to detect such - what are commonly termed - advanced persistent threats. A new approach is needed that takes the stepwise characteristics of this type of threats into account and links analysis methods to attack features. This paper takes up this challenge. First, an analysis framework is proposed to relate complex attack attributes to detection and business aspects. Second, the framework is used to define a development roadmap for designing advanced intrusion detection systems, such systems can analyze network traffic and client data at multiple network locations using both signature and anomaly detection methods derived from the intelligent data analysis field. Third, a test case is provided showing the potential power of the proposed development roadmap.

Trust metrics for survivable security systems

Critical information infrastructures contain specific nodes that provide security services; those are called security or trust centres. The goal of this paper is to find algorithms for enabling trust centres to become sustainable by sharing (superfluous) their security resources. The first algorithm indicates the number of segments the total distrusted infrastructure has to be split in to result in many smaller trusted groups. The second type of algorithms is meant to enable the autonomous trust valuation by the nodes. The third type of algorithm comprehends the creation of and the distribution of security assets to other resources.

Multi-level access control in the data pipeline of the international supply chain system

The Seamless Integrated Data Pipeline system was proposed to the European Union in order to overcome the information quality shortcomings of the current international supply chain information exchange systems. Next to identification and authorization of stakeholders, secure access control needs to be considered at design time of the new data pipeline system. This challenge is taken up in this paper. First, based on an analysis of the proposed data pipeline concept, access control requirements are being defined. Second, a new multi-level access control model is being designed. The resulting model organizes access control at two levels, namely, at country and at service level, herewith enabling secure information exchange between global stakeholders.

Towards a trustworthy short-range mobile payment system

In commercial clusters, like shopping malls, subways and university campuses, shops may act as service providers. Via proximity networks, they can provide services to users with short-range radio mobile devices, including financial services. Short-range mobile payment services represent therefore an alternative business opportunity for the financial industry. By enlisting these short-range services, users are not required anymore to have and to rely on cell phones only. In this paper, we propose a mobile payment model for any service cluster where users can clear their payments using short-range mobile devices only. Due to the possible lack of a (reliable) long range wireless communication network on the users side in this design, the service provider also acts as a payment gateway. The schema is designed to be trustworthy, since it includes a balanced trust model and an autonomous trust valuator.

Emergent information security in critical infrastructures

The growing tendency to inter-weaving infrastructures makes them more complex, less manageable, and more vulnerable to random system failures. Several information systems are usually added to manage and control such infrastructures. However, in this internet era these information infrastructures have increasingly become the targets of sophisticated (denial of service) attacks, which cause random failures. Common technologies designed to improve the resilience of critical nodes in such information infrastructures are based on costly, dedicated, and limited redundant hardware systems. This paper explores the possibility to apply self-organisation and resource sharing techniques between the security nodes to increase the infrastructure resilience. It is possible then to achieve perpetual availability without dedicated redundant systems.

Virtual Trust in Distributed Systems

This paper raises two problems of trusted services in distributed organizations. First, on a global scale trust becomes a hard issue to solve for many multinationals since there is no such thing as a global PKI, although many efforts try to overcome this gap. We propose an alternative non-institutionalized trust model to overcome this global trust dilemma. Second, trust prohibits real-time concurrent replication of the trusted service on redundants to increase dependability. We argue why the fuzzy concept trust does not permit replication techniques and propose an indirect approach to trust by indicators.

Dependable Security by Twisted Secret Sharing

Large scale networked information systems are referred to as critical information infrastructures when they provide critical services to the critical physical infrastructures. Critical information infrastructures contain specific nodes that provide security services, like authentication servers; those are called security or trust centres. The goal of this research is to fmd an algorithm for enabling those centres to become sustainable by sharing their (superfluous) security resources and to resist Byzantine failures. The proposed secret sharing algorithm takes care of allocating in advance the secret content of the suffering centre to other healthy centres, so that only an arbitrary majority of them can reconstruct the content. This perfect t,n- threshold scheme is suitable in dynamic networks as it has an adaptive access structure. It is compared to existing schemes rather simple as it is purely based on permutations. It is efficient, i.e. favourable information rate, as all shares are much shorter than the secret itself. Secondarily, each secret share is even additionally protected (encrypted) against the holder as well against any outsider.

Building a distributed security defence system

Given the complexity of infrastructures, current state of security technology and the limited budgets any security defense systems can be outnumbered by a sufficient number of random sequential failures, e.g. due to multiple DOS attacks. Complementary to the regular solutions where per node several identical dedicated nodes are added (i.e. redundants), a resource sharing approach between undedicated nodes is aimed to build a large scale cluster of redundants and approximate perpetual availability of security distributing nodes. In this work principles are acquired from related and unrelated fields to build a distributed defense system (DDS) that relies on resource sharing. The proposed protocol set, called Medusa, achieves this DDS by dissociating trust authority from identity and hardware, making trust a moveable emancipated commodity. As a moveable object trust can apply traditional fault tolerance techniques by process migration.

An Adaptive Trustworthiness Modelling Approach for Ubiquitous Software Systems

With the growing interest in ubiquitous applications, attention should be given to privacy and trust issues. Since these environments are quite dynamic, and interactive where devices possess low computing resources, new trust models are required that suit this environment and help overcome its limitations. In this paper, the authors explore the possibility to settle trust in an adaptive way prior to any information disclosure to enable adaptive privacy and security. The proposed model is simulated in a commercial cluster scenario where users can clear all their payments with their short-range ubiquitous devices. In this domain, the authors assume that the user has to rely on what they call payment disclosure service PDS to control his or her disclosure decisions. The proposed model respects the limitations of wireless devices such as limited connectivity and processing resources, while enabling autonomous and secure operations.

Trust algorithms in P2P file sharing networks

Many peer to peer (P2P) networks are used to share different types of files between users who are usually anonymous. To prevent the distribution of files that have been intentionally damaged, filled with false content or infected with malware, users need to be able to distinguish between trustworthy and malicious peers. One way to help a user in separating the good peers from the bad peers, is to provide him/her with a sophisticated trust algorithm that calculates the trustworthiness of any peer involved in a file sharing transaction. In this paper the results of a comparative study are presented as a first contribution. This study is based on taxonomy of 17 trust algorithms as found in literature. The evaluation is done by means of both a literature study and a series of simulation tests. By using the evaluation results, the advantages of each of the current algorithms are singled out, and next combined to develop and implement a new and better trust algorithm for file sharing in P2P networks. This is the second contribution.