Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Supporting differentiated classes of resilience in multilayer networks

Alashaikh, Abdulaziz (2017) Supporting differentiated classes of resilience in multilayer networks. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Download (8MB) | Preview

Abstract

Services provided over telecommunications networks typically have different resilience requirements and networks need to be able to support different levels of resilience in an efficient manner. This dissertation investigates the problem of supporting differentiated classes of resilience in multilayer networks, including the most stringent resilience class required by critical services. We incorporate an innovative technique of embedding a subnetwork, termed the spine, with comparatively higher availability values at the physical layer. The spine lays a foundation for differentiation between multiple classes of flows that can be leveraged to achieve both high resilience and differentiation. The aim of this research is mainly to explore, design, and evaluate the proposed spine concept model in multilayer networks. The dissertation has four major parts. First, we explore the spine concept through numerical analysis of simple topologies illustrating the potential benefits and the cost considerations of the spine. We develop heuristics algorithms to find suitable spines for a network based on the structural properties of the network topology. Second, an optimization problem is formulated to determine the spine. The problem encompasses estimates of link availability improvements, associated costs, and a total budget. Third, we propose a crosslayer mapping and spine-aware routing design problem with protection given mainly at the lower layer. The problem is designed to transfer lower layer differentiation capability to the upper layer network and flows. We provide two joint routing-mapping optimization formulations and evaluate their performance in a multilayer scenario. Fourth, the joint routing-mapping problem is redesigned with protection given in the upper network layer instead. This will create two isolated logical networks; one mapped to the spine and the other is mapped freely on the network. Flows are assigned a path or path-pair based on their class of resilience. This approach can provide more routing options yielding different availability levels. The joint routing-mapping design problems are formulated as Integer Linear Programming (ILP) models. The goal is to achieve a wider range of availability values across layers and high availability levels for mission-critical services without the need to use higher order protection configurations. The proposed models are evaluated with extensive numerical results using real network topologies.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Alashaikh, Abdulazizazizoozi@gmail.comasa440000-0002-1525-4500
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairTipper, Daviddtipper@pitt.edu
Committee MemberWeiss, Martinmbw@pitt.edu
Committee MemberKrishnamurthy, Prashantprashk@pitt.edu
Committee MemberGomes, Teresateresa@deec.uc.pt
Date: 22 September 2017
Date Type: Publication
Defense Date: 20 July 2017
Approval Date: 22 September 2017
Submission Date: 23 August 2017
Access Restriction: 2 year -- Restrict access to University of Pittsburgh for a period of 2 years.
Number of Pages: 186
Institution: University of Pittsburgh
Schools and Programs: School of Information Sciences > Telecommunications
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Differentiated Resilience, Availability, Multilayer Networks Design
Additional Information: This is the revised version of the dissertation.
Date Deposited: 01 Nov 2017 17:33
Last Modified: 22 Sep 2019 05:15
URI: http://d-scholarship.pitt.edu/id/eprint/33136

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item