Peter Komisarczuk

Identification of Malicious Web Pages with Static Heuristics

Malicious web pages that launch client-side attacks on web browsers have become an increasing problem in recent years. High-interaction client honeypots are security devices that can detect these malicious web pages on a network. However, high-interaction client honeypots are both resource-intensive and known to miss attacks. This paper presents a novel classification method for detecting malicious web pages that involves inspecting the underlying static attributes of the initial HTTP response and HTML code. Because malicious web pages import exploits from remote resources and hide exploit code, static attributes characterizing these actions can be used to identify a majority of malicious web pages. Combining high-interaction client honeypots and this new classification method into a hybrid system leads to significant performance improvements.

Review: Reinforcement learning for context awareness and intelligence in wireless networks: Review, new features and open issues

In wireless networks, context awareness and intelligence are capabilities that enable each host to observe, learn, and respond to its complex and dynamic operating environment in an efficient manner. These capabilities contrast with traditional approaches where each host adheres to a predefined set of rules, and responds accordingly. In recent years, context awareness and intelligence have gained tremendous popularity due to the substantial network-wide performance enhancement they have to offer. In this article, we advocate the use of reinforcement learning (RL) to achieve context awareness and intelligence. The RL approach has been applied in a variety of schemes such as routing, resource management and dynamic channel selection in wireless networks. Examples of wireless networks are mobile ad hoc networks, wireless sensor networks, cellular networks and cognitive radio networks. This article presents an overview of classical RL and three extensions, including events, rules and agent interaction and coordination, to wireless networks. We discuss how several wireless network schemes have been approached using RL to provide network performance enhancement, and also open issues associated with this approach. Throughout the paper, discussions are presented in a tutorial manner, and are related to existing work in order to establish a foundation for further research in this field, specifically, for the improvement of the RL approach in the context of wireless networking, for the improvement of the RL approach through the use of the extensions in existing schemes, as well as for the design and implementation of RL in new schemes.

Cognitive Radio-based Wireless Sensor Networks: Conceptual design and open issues

Traditional static spectrum allocation policies have been to grant each wireless service exclusive usage of certain frequency bands, leaving several spectrum bands unlicensed for industrial, scientific and medical purposes. The rapid proliferation of low-cost wireless applications in unlicensed spectrum bands has resulted in spectrum scarcity among those bands. Since most applications in Wireless Sensor Networks (WSNs) utilize the unlicensed spectrum, network-wide performance of WSNs will inevitably degrade as their popularity increases. Sharing of under-utilized licensed spectrum among unlicensed devices is a promising solution to the spectrum scarcity issue. Cognitive Radio (CR) is a new paradigm in wireless communication that allows sensor nodes as the unlicensed users or Secondary Users (SUs) to detect and use the under-utilized licensed spectrum temporarily. Given that the licensed or Primary Users (PUs) are oblivious to the presence of SUs, the SUs access the licensed spectrum opportunistically without interfering the PUs, while improving their own performance. In this paper, we propose an approach to build Cognitive Radio-based Wireless Sensor Networks (CR-WSNs). We believe that CR-WSN is the next-generation WSN. Realizing that both WSNs and CR present unique challenges to the design of CR-WSNs, we provide an overview and conceptual design of WSNs from the perspective of CR. The open issues are discussed to motivate new research interests in this field. We also present our method to achieving context-awareness and intelligence, which are the key components in CR networks, to address an open issue in CR-WSN.

A context-aware and Intelligent Dynamic Channel Selection scheme for cognitive radio networks

The tremendous growth in ubiquitous low-cost wireless applications that utilize the unlicensed spectrum bands has laid increasing stress on the limited and scarce radio spectrum resources. Given that the licensed or Primary Users (PUs) are oblivious to the presence of unlicensed or Secondary Users (SUs), Cognitive Radio (CR) is a new paradigm in wireless communication that allows the SUs to detect and use the underutilized licensed spectrums opportunistically and temporarily. In this paper, we propose a Context-aware and Intelligent Dynamic Channel Selection scheme that helps SUs to select channel adaptively for data transmission to enhance QoS, particularly throughput and delay. Our scheme is suitable for CR networks with mobile hosts. We formulate and design our scheme using Reinforcement Learning that offers a simple and yet practical solution. Channel heterogeneity, which is a feature unique to CR networks that has been ignored in previous studies, is considered in this paper. Simulation results reveal that the proposed scheme achieves very good performance.

Enhancing network performance in Distributed Cognitive Radio Networks using single-agent and multi-agent Reinforcement Learning

Cognitive Radio (CR) is a next-generation wireless communication system that enables unlicensed users to exploit underutilized licensed spectrum to optimize the utilization of the overall radio spectrum. A Distributed Cognitive Radio Network (DCRN) is a distributed wireless network established by a number of unlicensed users in the absence of fixed network infrastructure such as a base station. Context awareness and intelligence are the capabilities that enable each unlicensed user to observe and carry out its own action as part of the joint action on its operating environment for network-wide performance enhancement. These capabilities can be applied in various application schemes in CR networks such as Dynamic Channel Selection (DCS), congestion control, and scheduling. In this paper, we apply Reinforcement Learning (RL), including single-agent and multi-agent approaches, to achieve context awareness and intelligence. Firstly, we show that the RL approach achieves a joint action that provides better network-wide performance in respect to DCS in DCRNs. The multi-agent approach is shown to provide higher levels of stability compared to the single-agent approach. Secondly, we show that RL achieves high level of fairness. Thirdly, we show the effects of network density and various essential parameters in RL on the network-wide performance.

Applications of Reinforcement Learning to Cognitive Radio Networks

Cognitive Radio (CR) enables an unlicensed user to change its transmission and reception parameters adaptively according to spectrum availability in a wide range of licensed channels. The concept of a Cognition Cycle (CC) is the key element of CR to provide context awareness and intelligence so that each unlicensed user is able to observe and carry out an optimal action on its operating environment for performance enhancement. The CC can be applied in various application schemes in CR networks such as Dynamic Channel Selection (DCS), topology management, congestion control, and scheduling. In this paper, Reinforcement Learning (RL) is applied to implement the conceptual of the CC. We provide an extensive overview of our work including single-agent and multi-agent approaches to show that RL is a promising technique. Our contribution in this paper is to propose various application schemes using our RL approach to warrant further research on RL in CR networks.

High occupancy resource allocation for grid and cloud systems, a study with DRIVE

Economic models have long been advocated as a means of efficient resource allocation, however they are often criticized due to a lack of performance and high overheads. The widespread adoption of utility computing models as seen in commercial Cloud providers has re-motivated the need for economic allocation mechanisms. The aim of this work is to address some of the performance limitations of existing economic allocation models, by reducing the failure/reallocation rate, increasing occupancy and thereby increasing the obtainable utilization of the system. This paper is a study of high performance resource utilization strategies that can be employed in Grid and Cloud systems. In particular we have implemented and quantified the results for strategies including overbooking, advanced reservation, justin-time bidding and using substitute providers for service delivery. These strategies are analyzed in a meta-scheduling context using synthetic workloads derived from a production Grid trace to quantify the performance benefits obtained.

Identification of malicious web pages through analysis of underlying DNS and web server relationships

Malicious Web pages that launch drive-by-download attacks on Web browsers have increasingly become a problem in recent years. High-interaction client honeypots are security devices that can detect these malicious Web pages on a network. However, high-interaction client honeypots are both resource-intensive and unable to handle the increasing array of vulnerable clients. This paper presents a novel classification method for detecting malicious Web pages that involves inspecting the underlying server relationships. Because of the unique structure of malicious front-end Web pages and centralized exploit servers, merely counting the number of domain name extensions and Domain Name System (DNS) servers used to resolve the host names of all Web servers involved in rendering a page is sufficient to determine whether a Web page is malicious or benign, independent of the vulnerable Web browser targeted by these pages. Combining high-interaction client honeypots and this new classification method into a hybrid system leads to performance improvements.

An extensible, self contained, layered approach to context acquisition

Smart phones show increasing capabilities for context-aware applications. The development of such applications involves implementation of mechanisms for context acquisition and context adaptation. To facilitate efficient use of the devices resources and avoid monitoring the same context changes from multiple points, it is necessary that applications share the context acquisition mechanism. In this paper, we intend to develop a generic context acquisition engine which is capable of context capturing, composition and broadcasting. By deploying the engine on a mobile device, context changes are monitored from single point and disseminated to various context aware applications running on the same device. As a proof of concept, the context acquisition engine has been implemented on the Android platform.

On Multi-Channel MAC Protocols in Cognitive Radio Networks

Cognitive Radio (CR) exploits underutilized licensed spectrums to improve its bandwidth availability. Using CR technology, a node is able to adapt its transmission and reception radio parameters including channel frequency dynamically according to local spectrum availability. For channel access between wireless nodes, a cognitive Medium Access Control (MAC) protocol is necessary to coordinate the CRs. Multi-channel MAC protocol extensions have been proposed in IEEE802.11 to enable a node to operate in multiple channels in order to improve network-wide throughput. These multi-channel MAC protocols have several functions that can be leveraged by cognitive MAC protocols due to their similarities in certain aspects, though the CR has an additional requirement to cope with the existence of licensed users that have higher authority over the channels. Current research in cognitive MAC protocols assumes the availability of a common control channel at all times, which is an approach in the multi-channel MAC protocols. This approach has certain hardware requirements that may not be readily available at CR nodes. Hence, other approached may be necessary. In this paper, various types of multi-channel MAC protocols are reviewed, followed by discussion of their merits and demerits in multi-channel environments. The purpose is to show the additional functionalities and challenges that each multi-channel MAC protocol has to offer and address in order to operate in multihop CR networks. By providing discussion on possible technology leverage from multi-channel to cognitive MAC protocols, we aim to establish a foundation for further research and discussion.