There are two main challenges for flexible service provisioning on top of today's Internet. First the end-to-end networking paradigm which is based on a service and data unaware transport, and secondly a lack of high level communication abstractions.With this thesis we approach both challenges. To address the first point we propose a methodology to integrate the Network Provider into the process of service provisioning. The introduced cooperative service provisioning (CSP) enables an interaction between Service Provider, Network Provider and Client. CSP can be applied in classical Provider as well as novel Peerto-Peer scenarios.To address the second point, we adopt a Service Overlay approach. A new method based on the extension of Distributed Hash Tables (DHTs) is proposed to realise a distributed control plane for a decentral and reactive creation of Service Overlays. To archive this, a methodology to reduce Service Overlay related routing problems to search problems was developed. With regard to the required Quality of Service (QoS) measurements, the thesis focuses on scalable and time efficient estimation techniques. As a result, two new delay prediction schemes have been developed which improve estimation accuracy when compared to the state of the art.Furthermore, a novel approach to bandwidth estimation is investigated. It combines methods from landmark based delay estimation with a transformation step and has favourable estimation accuracy.
Recent platforms for Peer-to-Peer (P2P) based video distribution such as Zattoo and Joost are based on End-Host or Application Layer Multicast techniques. An on demand adaptation of audio/video data to end user terminals or to the transport and error characteristics of the client access technology is not supported. To address such issues in the scale of P2P Networks, we describe a cooperative service provisioning principle (CSP) based on the assumption that selected peers offer processing services to the community. In contrast to recent approaches our main focus in this paper is on the applicability of QoS estimation techniques to decrease time and measurement complexity of CSP. We provide a state of the art in landmark-based delay estimation and introduce a novel technique for landmark-based bottleneck bandwidth prediction showing favourable results for networks with widest path routing
In recent years, peer-to-peer overlay networks have become a popular communication paradigm with the potential to further change communication fundamentally in the future. Overlays allow communication abstraction but suffer from one inherent problem: The overlay is unaware of the context of a service or the context of a service consumer. The concept of context-awareness emerged out of the research done within the area of ubiquitous computing. Context-aware computing is one key technology to enable services and applications in the communication environment to adapt their behaviour based on the knowledge of environmental (contextual) information, thereby enhancing the system’s ability to become ever more responsive to the needs of the end-user or application domain. In this chapter we first introduce context and context architectures in general. In the remainder of the chapter we focus on the question: How can highly distributed context information be located and retrieved regarding small-scale as well as large-scale networks, addressing the topics of inter-domain management and scalability of context architectures?
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.