Technological and Economic Aspects for Quality of Service in Multi-Domain Alliances
Isabel Amigo
PhD thesis from Universidad de la Repúbica. Facultad de Ingeniería. IIE - 2013
Advisor: Pablo Belzarena
Co-advisor: Sandrine Vaton
Research Group(s): Analisis de Redes, Trafico y Estadisticas de Servi (art)
Department(s): Telecomunicaciones
Descargar la publicación : tesisIAmigo.pdf [2.2Mo]  


Providing end-to-end quality-assured services implies many challeng es, which go beyond technical ones, involving as well economic and even cultural or political issues. In this thesis we first focus on a technical problem and then intent a more holistic regard to the who le problem, considering at the same time Network Service Providers (NSPs), stakeholders and buyers behaviour and satisfaction. One of the most important problems when deploying interdomain path selection with Quality of Service (QoS) requirements is being able to rely the computations on metrics that hold for a long period of time. Our proposal for solving that problem is to compute bounds on the metrics, taking into account the uncertainty on the traffic demands. In particular , we will explore the computation of the maximum end-to-end delay of traversing a domain considering that the traffic is unknown but bounded. Since this provides a robust QoS value for traversing the NSP or Autonomous System (AS), without revealing confidential information, we claim that the b ound can be safely conceived as a metric to be announced by each AS to the entities performing the path selection, in the process of interdomain path selection. We show how the maximum dela y value is obtained for an interdomain bandwidth demand and we propose an exact method and a numerical approximation method for computing it, neither of which rely on a complex monitoring infrastructure. Simulations with real data that illustrate the problem and validate our results are also presented. In the multidomain context economics and policies become more complex. In this regard, AS alliances or federations are envisaged to emerge in the near future as a means of selling end-to-end quality-assured services through interdomain networks. This collaborative paradigm mainly responds to the ever increasing Internet traffic volumes that requires assured quality, and constitutes a new business opportunity for NSPs. However, current Internet business rules are not likely to satisfy all involved partners in this emerging scenario. How the revenue is shared among NSPs must be agreed in advance, and should enforce economic incentives to join an alliance and remain in it, so that the alliance remains stable. Inspired by this scenario, we propose a complete framework for selling interdomain quality- assured services, and subsequently distributing revenues, in an AS alliance context. We state the problem as a network utility maximization problem with QoS constraints and show that a distributed solution can be carried out. With respect to the revenue sharing problem, we formally formulate the properties the revenue sharing method should fulfil and argue why the existing methods are not suitable. We propose a family of solutions to the revenue sharing problem such that the economic stability and efficiency of the alliance in the long term is guaranteed. The proposed method is based on solving a series of Optimization Problems and considering statistics on the incomes. We then move to a more holistic approach and consider the interactio ns with the monitoring plane and the buyers behaviour. We propose a simple pricing scheme and study it in detail, in order to use QoS monitoring information as feedback to the business plane, with the ultimate objective of improving the sellers revenue. In our framework, ass ured-quality Services are sold through first-price auctions, and in case of failure, a percentage of the price paid for the service is given back to the buyers. We derive the expression for the willingnes s to pay and we model the pricing problem through a Stackelberg game. We solve the game to sh ow that the equilibrium that maximizes the sellers revenue implies reimbursing 100% in case of failures. The previous study is built upon a strong symmetry hypothesis, that is, buyers are assumed to be symmetric, and so do services in sale. In order to relax this assumption we present a simulative approach and evaluate the proposed pricing scheme with its aid. Results of simulations are shown in different scenarios, which in particular have shown results that are coherent with the analytical ones. That is to say, that in the evaluated scenarios reimbursing 100% provides more revenue to the seller than when no reimbursement is in place

Datos adicionales


Referencias BibTex

Descargar BibTex bibtex

Otras publicaciones

» Isabel Amigo