Ecole d'ingénieur et centre de recherche en télécommunications

Game theoretic analysis of distributed systems : design and incentives

Toka, Laszlo

Thesis

This dissertation studies incentive aspects of distributed systems in which limited private or public resources must be allocated among selfish autonomic participants. Our goal is to design mechanisms which ensure the efficiency and fairness of resource allocation in such systems. We employ selfish user models and we investigate the results of our proposed schemes. We also design distributed optimization algorithms intended for practical implementation. First, we target backup services in peer-to-peer systems, i.e., distributed networks of functionally equal peers, where users save their backup data on the underutilized storage devices of one another over the Internet. As a main characteristic, no scalability problems arise since more users provide larger overall storage space and bandwidth. Furthermore, spatial and ownership diversity of storage hosts assure the availability of backed up data. However, the management of users not willing to share their local resources with other participants is extremely important to keep the system operational. Moreover, ensuring high quality of service in such a peer-to-peer network requires careful system design. Our novel data redundancy and peer selection policies provide reliable backup service in return for fair resource contribution of the users. Second, we examine the potential of a dynamic spectrum management framework that enables sequential allocation of frequency bands for wireless service providers. We present our distributed system design on allocation and pricing with the goal of achieving efficient spectrum utilization, flexible allocations and incentive-compatibility, considering physical interference among frequency licensees. Our work provides insights on emerging optimization problems related to the allocation. We suggest heuristic solutions to these problems based on our analytic results. We evaluate the proposed framework and algorithms numerically, and we conclude that our proposed heuristics can be the cornerstones of a flexible distributed dynamic allocation system.

Document Bibtex

Type:Thèse
Langue:English
Ville:
Date:
Département:Réseaux et Sécurité
Eurecom ref:3310
Copyright: © TELECOM ParisTech. Personal use of this material is permitted. The definitive version of this paper was published in Thesis and is available at :
Bibtex: @phdthesis{EURECOM+3310, year = {2011}, title = {{G}ame theoretic analysis of distributed systems : design and incentives}, author = {{T}oka, {L}aszlo}, school = {{T}hesis}, month = {01}, url = {http://www.eurecom.fr/publication/3310} }