Steven Simpson

Scalable adaptive hierarchical clustering

We propose a new application-level clustering algorithm capable of building an overlay spanning tree among participants of large multicast sessions, without any specific help from the network routers. The algorithm and associated protocol are shown to exhibit scalable properties.

Quality evaluation in peer-to-peer IPTV services

Modern IPTV services are comprised of multiple comprehensive service elements in the entire content delivery chain to maximise the efficiency in networking. Audio-visual content may experience various types of impairments during content ingest, processing, distribution and reception. While some impairments do not cause noticeable distortions to the delivered content, many others such as the network transmission loss can be highly detrimental to the user experience in content consumption. In order to optimise service quality and to provide a benchmarking platform to evaluate the designs for future audio-visual content distribution system, a quality evaluation framework is essential. We introduce such an evaluation framework to assess video service with respect of user perception, while supporting service diagnosis to identify root-causes of any detected quality degradation. Compared with existing QoE frameworks, our solution offers an advanced but practical design for the real-time analysis of IPTV services in multiple service layers.

A framework for resilience management in the cloud

Cloud environments make resilience more challenging because of the sharing of non-virtualised resources, frequent reconfigurations, and cyber attacks on these flexible and dynamic systems. We present a Cloud Resilience Management Framework (CRMF), which models and then applies an existing resilience strategy in a cloud operating context to diagnose anomalies. The framework uses an end-to-end feedback loop that allows remediation to be integrated with the existing cloud management systems. We demonstrate the applicability of the framework with a use-case for effective cloud resilience management.ZusammenfassungCloud-Umgebungen stellen wegen der gemeinsamen Nutzung von nicht-virtualisierten Ressourcen, häufiger Rekonfigurationen und Cyber-Angriffen auf diese flexiblen und dynamischen Systeme größere Herausforderungen an Ausfallsicherheit. In dieser Arbeit wird ein Cloud Resilience Management Framework (CRMF) präsentiert, das eine bereits existierende Ausfallsicherheitsstrategie im Kontext eines Cloudbetriebs modelliert und dort anwendet, um Anomalien zu erkennen. Das Framework benutzt eine Ende-zu-Ende-Feedbackschleife, die es ermöglicht, Problembehebung in vorhandene Cloud-Managementsysteme zu integrieren. Weiterhin wird die Anwendbarkeit dieses Frameworks durch einen Anwendungsfall mit effizientem Cloud Resilience Management gezeigt.

Component Selection for Heterogeneous Active Networking

Active Networking (AN) involves the processing of programs in heterogeneous networking environments. There are severalA N solutions, exposing different APIs and using different languages, and each may be appropriate for different tasks such as high-speed multimedia processing or low-speed routing adjustments.We describe our active node system, LANode, that separates control- and data-plane activities, and introduce Component Compatibility Markup Language (CCML), a critical component of LANode that allows it to be applied to heterogeneous platforms.

Improving Wireless Sensor Network Resilience with the INTERSECTION Framework

This paper investigates the INTERSECTION Framework’s ability to build resilience into Wireless Sensor Networks. The framework’s general details are examined along with general approaches to protection against misbehaving WSN nodes. Three different approaches to remediation were implemented and tested on a TinyOS network for their different operating characteristics.

User-level fairness delivered: Network resource allocation for adaptive video streaming

HTTP adaptive streaming (HAS) technology is becoming a popular vehicle for online video delivery. HAS applications often compete for network resources without any coordination between each other in a shared network. This leads to quality of experience (QoE) fluctuations and unfairness between end users. This paper introduces a user-level fairness model (UF) which exploits video quality, switching impact and cost efficiency as the fairness metrics to achieve user-level fairness in resource allocation. Experimental results demonstrate how this model is a foundation to orchestrate the resource consumption of HAS streams.

Storisphere: from TV watching to community story telling

Conventional television services have been increasingly challenged by the more interactive and user-centric video sharing applications. With the growing popularity of social networks and video services, users are becoming the editors and broadcasters of their own stories. User-generated video content, which provides unique perspectives from individuals, is likely to be the new medium to complement professional broadcast TV for story sharing, especially in user communities of specific interest. We have developed Storisphere to provide a web-based collaborative video content workspace for members of a community to compose and share video stories, using desktop or mobile devices. Storisphere is currently being evaluated for video storytelling by various user communities.

Identifying Legitimate Clients under Distributed Denial-of-Service Attacks

Distributed Denial of Service (DDoS) attacks are a persistent, current, and very real threat to networks. Expanding upon a flexible distributed framework for network remediation utilising multiple strategies, we examine a novel fusion of methods to maximise throughput from legitimate clients and minimise the impact from attackers. The basic approach is to build up a whitelist of likely legitimate clients by observing outgoing traffic, presenting a challenge though proof-of-work, and providing flow cookies. Traffic that does not match the expected profile is likely attack traffic, and can be heavily filtered during attack conditions. After we incrementally develop this approach, we explore the positive and negative impacts of this approach upon the network and analyse potential countermeasures.

A Semi-Autonomic Framework for Intrusion Tolerance in Heterogeneous Networks

A suitable strategy for network intrusion tolerance-- detecting intrusions and remedying them--depends on aspects of the domain being protected, such as the kinds of intrusion faced, the resources available for monitoring and remediation, and the level at which automated remediation can be carried out. The decision to remediate autonomically will have to consider the relative costs of performing a potentially disruptive remedy in the wrong circumstances and leaving it up to a slow, but more accurate, human operator. Autonomic remediation also needs to be withdrawn at some point --- a phase of recovery to the normal network state. In this paper, we present a framework for deploying domain-adaptable intrusion-tolerance strategies in heterogeneous networks. Functionality is divided into that which is fixed by the domain and that which should adapt, in order to cope with heterogeneity. The interactions between detection and remediation are considered in order to make a stable recovery decision. We also present a model for combining diverse sources of monitoring to improve accurate decision making, an important pre-requisite to automated remediation.

REDO RSVP: Efficient Signalling for Multimedia in the Internet

Alarming reports of performance and scalability problems associated with per-flow reservations have led many to doubt the ultimate utility of RSVP and the Integrated Services Architecture that relies on it. Because we axe convinced of the need for some form of resource reservation, to support multimedia communications in the Internet, we have set about trying to improve RSVP. By careful study of the protocol, we have identified areas for improvement, and propose REDO RSVP, a reduced overhead version that includes a fast establishment mechanism (FEM). In this paper we describe the rationale for REDO RSVP and present a detailed analysis of its features and operations. We also analyse REDO RSVP by means of simulations, and show that it offers improvements to the performance of RSVP.