Cost Effective and Survivable Wide-area Topology of Telecommunication Cabling

Project: Research

View graph of relations

Researcher(s)

Description

In several recent disasters, it has become apparent that submarine communication cables havebecome the Achilles’ heel of the Internet. In our region alone (Hong Kong/Taiwan/Japan) aswell as in other places in the world, many undersea cables are in close proximity of eachother, and located in earthquake-prone areas. This points to the importance of understandingthe consequences of such topology designs from the perspective of disaster mitigation, and tothe need of developing methodologies to improve network survivability.An important measure considered in this project is the network survival probability definedas the probability that the network remains connected in a disaster scenario. We address thefundamentally important problem of how to optimize cabling between a given set of nodespositioned on the surface of the Earth. We aim at either minimizing cost subject tomaintaining the network survival probability above a required level, or alternatively,maximizing the network survival probability subject to cost constraints. To provide acomplete survivable network design solution, we will also integrate our methodology withconventional designs that provide capacity assignment subject to isolated failures.Our approach is to develop a joint probability function over a sphere (modeling the Earth’ssurface) of the location of a disaster (normally an earthquake) and of its effect on thetelecommunication infrastructure. This probability function should be consistent with knownearthquake engineering data, and yet sufficiently simple to lead to an approximate analyticalexpression for the network survival probability. In addition to analyses yielding exactsolutions, Monte Carlo simulations will be used to perform comprehensive simulation studiesof realistic scenarios based on real seismic data and to validate the analytical solutions.This project falls under the field of disaster survivability, a topic which has recently gainedpopularity in the context of telecommunications network design. This proposal allows the useof curved cables for cost effectiveness with guaranteed disaster survivability. Unlike existingwork in the literature, our research is the first of its kind that deals with optimizing networktopologies in the context of a disaster scenario where links between adjacent nodes are notnecessarily straight lines.The team participating in this collaborative project includes telecommunications researchers,earthquake engineers, and mathematicians. Through the proposed collaboration,mathematical techniques and engineering knowledge will be combined towards solving animportant real life problem which requires an urgent solution as we seek to save the Internetfrom its Achilles’ heel.

Detail(s)

Project number8730032
Grant typeCRF
StatusFinished
Effective start/end date1/06/1430/05/18