Abstract

Our research goal is to investigate the FTTX (Fiber-to-the Home/Premises/Curb) passive optical network (PON) for the deployment of BISAN (Broadband Internet Subscriber Access Network) to exploit the opportunities of optical fiber enabled technologies as well as of passive switching equipment. Indeed, the deployment of FTTX PON is the most
OPEX-friendly scenario, because it allows for completely passive access networks through minimizing the number of active components in the network. Previously, most FTTX PON
architectures are designed based on the principle of either time division multiplexing (TDM) technology or wavelength division multiplexing (WDM) technology. We focus on designing the best possible architectures of FTTX PON, specifically hybrid PONs, which embraces both TDM and WDM technology. A hybrid PON architecture is very efficient as it is not limited to any specific PON technology, rather it is flexible enough to deploy TDM/WDM technology depending on the type (i.e unicast/multicast) and amount of traffic demand of the end-users. The advantages of a hybrid PON are of two folds: (i) it can offer increased data rate to each user by employing WDM technology, (ii) it can provide flexible bandwidth utilization by employing TDM technology.
In this thesis, we concentrate on determining the optimized covering of a geographical area by a set of cost-effective hybrid PONs. We also focus on the greenfield deployment of a
single hybrid PON. It should be worthy to mention that while investigating the deployment of hybrid PONs, the research community around the world considers the specifications of either the physical layer or the optical layer. But an efficient planning for PON deployment
should take into account the constraints of the physical and optical layers in order that both layers can work together harmoniously. We concentrate our research on the network dimensioning and the selection as well as the placement of the switching equipment in hybrid PONs with the intention of considering the constraints of both physical and optical layers.
We determine the layout of an optimized PON architecture while provisioning wavelengths in a hybrid PON. We also propose to select the switching equipment depending on the type (unicast/multicast) of traffic demand. Finally, we determine the best set of hybrid PONs along with their cascading architecture, type and location of their switching equipment while satisfying the network design constraints such as the number of output ports of the
switching equipment and maximum allowed signal power loss experienced at each end user’s premises.
In this thesis, we propose two novel schemes for the greenfield deployment of a single hybrid PON. The first scheme consists of two phases in which a heuristic algorithm and a novel column generation (CG) based integer linear programming (ILP) optimization model are proposed in the 1st and 2nd phase respectively. In the second scheme, a novel integrated CG based ILP cross layer optimization model is proposed for the designing of a single hybrid
PON.
We also propose two novel schemes to deal with the greenfield deployment of multiple hybrid PONs in a given geographical area. These two schemes determine the best set of cost-effective hybrid PONs in order to serve all the end users in a given neighborhood.
The first scheme executes in four phases in which two heuristic algorithms, a CG based ILP model and an ILP optimization model are proposed in the 1st, 2nd, 3rd and 4th phase respectively. In the second scheme, an ILP model as well as a CG based ILP model, another ILP model as well as another CG based ILP model, a CG based ILP model and an ILP optimization model are proposed during four consecutive phases.
Our proposed scheme can optimize the design of a set of hybrid PONs covering a given geographic area as well as the selection of the best cascading architecture 1/2/mixedstage)
for each selected PON. It minimizes the overall network deployment cost based on the location of the OLT and the ONUs while granting all traffic demands. The scheme
emphasizes on the optimum placement of equipment in a hybrid PON infrastructure due to the critical dependency between the network performances and a proper deployment of
its equipment, which, in turn depends on the locations of the users. It is a quite powerful scheme as it can handle data instances with up to several thousands ONUs. On the basis of the computational results, the proposed scheme leads to an efficient automated tool for network design, planning, and performance evaluation which can be beneficial for the network designers.