Requirements and alternatives for Carrier Ethernet Transport based aggregation networks in a country-wide governmental network

OData support
Jakab Tivadar
Department of Networked Systems and Services

The motivation of my final thesis is to connect settlements of a sub-regional to the backbone network, which covers the sub-regional centers. The sub-regional settlements are connected with aggregation network segment to the government’s backbone network. The Carrier Ethernet Transport (CET) technology offers wide and range of new solutions and services in the aggregation and backbone networks. The task is to specify and evaluate realistic network scenarios using adaptable scaling models with available CET ring structure for aggregation networks to extend a Hungarian hypothetical IP/MPLS-WDM government communication backbone with proper aggregation. The government backbone network connects the cities of the 174 sub-regional centers, however the settlements of each sub-regional should be served by the CET aggregation networks. The thesis aims to introduce simple models, methods and solutions applicable to design and plan CET aggregation network for this purpose.

In the thesis first I present the progress of Ethernet technology, how can we use it in a service provider’s environment. I review the technology gaps and the solutions for the problems.

In the following chapter I describe traffic models, which can derive traffic flows generated in a sub-regional town. Demographic, administrative, health, education and cultural statics- and data, as well as assumed typical applications serve as basics of simple models used to estimate bandwidth requirements of settlements connected to the network. I represent the traffic demands in a database, which is used later in the design of Ethernet rings. I defined some classes of services, according to the protection requirements, i.e. which application data may not be compromised, as at highest priority service class.

The next phase I present the aggregation network design. First, I expound a mathematical linear modeling description. The developed linear programming (LP) model is able to determine the shortest trail ring topologies in a graph that cover the settlements to be connected to a given small regional center. The model is focused on the cable topology, I assumed cables or potential cable trails next to railroads, highways and primary roads. Second, I present the system dimensioning, based on the Ethernet ring design models. The goal is to connect the rings to backbone network, the technical problem the ring stacking problem i.e. how many ring system required to serve the bandwidth requirements of the connected settlements, and how the settlements are assigned to the different ring systems. Some detailed case studies are presented to illustrate the resource needs of different protection options. Through my completed work I developed network models in linear programming environment. The available free LP_Solve solver engine is used to evaluate the linear programming models. Two sub-regionals of different economic development levels and geographical structures were chosen for detailed case studies. Case studies cover the calculation of proper cable topology, as well as the dimensioning and configuring of aggregation rings.

At the end of the study I proposed a simple cost estimation model and specified some problems and task for potential further work on the topic.


Please sign in to download the files of this thesis.