Jodie Wetherall

A Cluster-Based Approach to Consensus Based Distributed Task Allocation

This paper presents a novel extension to the Consensus-Based Bundle Algorithm (CBBA), which we have named Cluster-Formed Consensus-Based Bundle Algorithm (CFCBBA). CF-CBBA is designed to reduce the amount of communication required to complete a distributed task allocation process, by partitioning the problem and processing it in parallel clusters. CF-CBBA has been shown, in comparison with baseline CBBA, to require less communication when allocating tasks. Three key aspects of task allocation have been investigated, (a) the time taken to allocate tasks, (b) the amount of communication necessary to satisfy the requirements of distributed task allocation algorithms such as CBBA, and (c) the efficiency with which a collection of tasks (a mission) is completed by a group of robots (a collective).

A large-scale zero-day worm simulator for cyber-epidemiological analysis

The cost of a single zero-day network worm outbreak has been estimated at US

SUPERMAN: Security Using Pre-Existing Routing for Mobile Ad hoc Networks

2.6 billion. Additionally zero-day worm outbreaks have been observed to spread at a significant pace across the global Internet, with an observed infection proportion of more than 90 percent of vulnerable hosts within 10 minutes. The threat posed by such fast-spreading malware is therefore significant, particularly given the fact that network operator / administrator intervention is not likely to take effect within the typical epidemiological timescale of such infections. An accepted tool that is used in researching the threat presented by zero-day worms is the use of simulation systems. However when considering zero-day worm outbreaks on the Internet there are persistent issues of scale and fidelity. The Internet Worm Simulator (IWS) reported in this paper is designed to address these issues by presenting a novel simulation method that, on a single workstation, can simulate an entire IPv4 address space on a node-by-node basis. Being able to simulate such a large-scale network enables the further analysis of characteristics identified from worm analysis. As IWS does not rely on mathematical approximation, the epidemiological attributes identified from real-world data can be tested for zero-day worm outbreaks on the Internet. Experimentation indicates that IWS is able to accurately simulate and corroborate with reported characteristics of two previous zero-day worm outbreaks. It is intended that, in future, IWS may be used to aid both in the analysis of previous worm outbreaks and the testing of hypothetical zero-day worm outbreak scenarios.

Virtual closed networks: A secure approach to autonomous mobile ad hoc networks

The flexibility and mobility of Mobile Ad hoc Networks (MANETs) have made them increasingly popular in a wide range of use cases. To protect these networks, security protocols have been developed to protect routing and application data. However, these protocols only protect routes or communication, not both. Both secure routing and communication security protocols must be implemented to provide full protection. The use of communication security protocols originally developed for wireline and WiFi networks can also place a heavy burden on the limited network resources of a MANET. To address these issues, a novel secure framework (SUPERMAN) is proposed. The framework is designed to allow existing network and routing protocols to perform their functions, whilst providing node authentication, access control, and communication security mechanisms. This paper presents a novel security framework for MANETs, SUPERMAN. Simulation results comparing SUPERMAN with IPsec, SAODV, and SOLSR are provided to demonstrate the proposed frameworks suitability for wireless communication security.

Simulation of zero-day worm epidemiology in the dynamic, heterogeneous Internet

The increasing autonomy of Mobile Ad Hoc Networks (MANETs) has enabled a great many large-scale unguided missions, such as agricultural planning, conservation and similar surveying tasks. Commercial and military institutions have expressed great interest in such ventures; raising the question of security as the application of such systems in potentially hostile environments becomes a desired function of such networks. Preventing theft, disruption or destruction of such MANETs through cyber-attacks has become a focus for many researchers as a result. Virtual Private Networks (VPNs) have been shown to enhance the security of Mobile Ad hoc Networks (MANETs), at a high cost in network resources during the setup of secure tunnels. VPNs do not normally support broadcast communication, reducing their effectiveness in high-traffic MANETs, which have many broadcast communication requirements. To support routing, broadcast updates and efficient MANET communication, a Virtual Closed Network (VCN) architecture is proposed. By supporting private, secure communication in unicast, multicast and broadcast modes, VCNs provide an efficient alternative to VPNs when securing MANETs. Comparative analysis of the set-up overheads of VCN and VPN approaches is provided between OpenVPN, IPsec, Virtual Private LAN Service (VPLS), and the proposed VCN solution: Security Using Pre-Existing Routing for MANETs (SUPERMAN).

Comparative Performance of a Parallel Implementation of an Internet-Scale Zero-Day Worm Epidemiology Simulator

The cost of a single zero-day network worm outbreak on the global Internet has been estimated at US

Investigation into the effects of transmission-channel fidelity loss in RGBD sensor data for SLAM

2.6 billion. In addition, zero-day network worm outbreaks have been observed that spread at a significant pace across the Internet, with an observed infection proportion of more than 90% of vulnerable hosts within 10 minutes. The threat posed by such fast-spreading malware to defence systems and national security is therefore significant, particularly given the fact that network operator/administrator intervention is not likely to take effect within the typical epidemiological timescale of such infections. An accepted technology that is used to research the security threat presented by zero-day worms is that of simulation systems; however, only a subset of these focus on the Internet and issues persist regarding how representative these are of the Internet. The design of a novel simulator developed to address these issues, the Internet Worm Simulator (IWS), is presented along with experimental results for a selection of previous worm outbreaks compared against observed, empirical data and hypothetical outbreak scenarios. Based on a finite state machine for each network host, the IWS incorporates the dynamic, heterogeneous characteristics of the Internet and, on a single workstation, is able to simulate an IPv4-sized network. Based on the analysis presented, the authors conclude that the IWS has the capability to simulate zero-day worm epidemiology on the dynamic, heterogeneous Internet for a variety of scenarios. These include simulating previous worm outbreaks that demonstrate random-scanning and hit list behaviour, as well as hypothetical scenarios that include a large susceptible populous and stealth-like behaviour.

A survey on evolutionary-aided design in robotics

The threat posed by fast-spreading malware is significant, particularly given the fact that network operator/administrator intervention is not likely to take effect within the typical epidemiological timescale of such infections. The cost of zero-day network worm outbreaks has been estimated to be up to US

Investigating an A-star algorithm-based fitness function for mobile robot evolution

2.6 billion for a single worm outbreak. Zero-day network worm outbreaks have been observed that spread at a significant pace across the global Internet, with an observed rate of reaching more than 90 percent of vulnerable hosts within 10 minutes. An accepted technology that is used in addressing the security threat presented by zero-day worms is the use of simulation systems, and a common factor determining their efficacy is their performance. An empirical comparison of a sequential and parallel implementation of a novel simulator, the Internet Worm Simulator (IWS), is presented detailing the impact of a selection of parameters on its performance. Experimentation demonstrates that IWS has the capability to simulate up to 91.8 million packets transmitted per second (PTS) for an IPv4 address space simulation on a single workstation computer, comparing favourably to previously reported metrics. It is concluded that in addition to comparing PTS performance, simulation requirements should be taken into consideration when assessing the performance of such simulators.

Securing autonomous networks through virtual closure

Simultaneous Location and Mapping (SLAM) is computationally expensive, and requires high-fidelity sensor data. This paper investigates the effects of transmission channel fidelity loss in Red-Green-Blue-Depth (RGBD) sensor data. A mobile robotic platform developed for Explosive Ordinance Disposal (EOD) is used, with a highly constrained data and video link to a base station which computes a SLAM solution. Experiments were conducted offline, using well known data-sets with ground truth data, and their results have been compared to determine the effect of fidelity loss under various multiplexing approaches with a constrained transmission channel.