טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
Ph.D Thesis
Ph.D StudentYallouz Jose
SubjectSchemes for Network Survivability in QoS-Supporting
Architectures
DepartmentDepartment of Electrical Engineering
Supervisor Professor Ariel Orda
Full Thesis textFull thesis text - English Version


Abstract

The Internet is the cornerstone of the last century digital revolution, which has become an integral part of our social lives and turned into an extensive force of the current world. However, the rapid evolution of this world-wide computer network has largely disregarded two essential aspects, namely Quality of Service (QoS) and survivability. This dissertation establishes a comprehensive methodology for efficiently coping with these two important aspects.


Network survivability, i.e. the resistance of the network against failures, has been acknowledged as an issue of major importance for the network infrastructure. In this study, we develop a novel survivability concept, termed tunable survivability, which offers major improvements over traditional approaches. While the traditional approach aims to provide full (100%) protection against network failures through disjoint paths, tunable survivability offers a quantitative measure for specifying the desired level (0%-100%) of survivability. We study this concept for two well-known transmission methods, namely unicast and broadcast. Moreover, we establish efficient algorithmic schemes for optimizing the level of survivability under QoS bounds, representing either additive or bottleneck metrics. Another important contribution is the extension of a well-known result for finding minimum weight link-disjoint paths by tuning the number of its common nodes.



Quality of Service (QoS), i.e., the capability of a network to provide guarantees to deliver predictable results, has always been a central topic of networking research. Besides the consideration of QoS in respect of tunable survivability, our study focuses on the evaluation of QoS guarantees of a specific routing unit, namely Stackable Router. Furthermore, we introduce and investigate a new QoS metric that measures the number of shortest path covered by a spanning tree.