An optical network provides a typicalinfrastructure over that a verity of services may be delivered. These networksare also capable of delivering bandwidth in a versatile manner, supportscapability up gradation and transparency in information transmission. Itconsists of optical supply (LED, LASER) as transmitter and optical fibre astransmission medium with alternative connectors and photo detector, receiverset. However owing to limitation of electronic processing speed, it’s impossibleto use all the BW of an optical fibre employing a single high capacity channelor wavelength. The first drawback in a very WDM network design is to search outthe simplest attainable path between a source-destination node pair and assignavailable wavelength to this path for data transmission. To see the simplestpath a series of measurements are performed which that are refereed asperformance matrices. From these performance matrices, the Quality of Serviceparameters are determined.
In this digital era the communication demand has increased from previous eras due to introduction of new communication techniques. As we can see there is increase in clients day byday, so we need huge bandwidth and high speed networks to deliver good quality of service to clients. Fiber optics communication is one of the major communication systems in modern era, which meets up the above challenges. This utilizes different types of multiplexing techniques to maintain good quality ofservice without traffic, less complicatedinstruments with good utilization ofavailable resources .Wavelength Division Multiplexing (WDM) is one of them with good efficiency. It is based on dynamic light-path allocation. Here we have to take into consideration the physical topology of the WDM network and the traffic. We have taken performance analysis as parameter to analyse which type of topology is best suited to implement in real life application without degrading quality of service (QoS).
Depending on the type of theoptical network considered, this problem may require the consideration ofvarious restrictions on the physical and logical topology. The restrictionsrelated to the physical topology of WDM optical networks usually depend on thetype of node equipment used and the capacity of the optical fibers available inthe network. The WDM optical network design problem studied in this paper dealswith an all-optical (also referred to as transparent) architecture. Currenttrends in optical networks indicate they are gradually moving from an opaquearchitecture (i.e., an architecture where the optical signal carrying trafficundergoes an optical-electronicoptical (OEO) conversion at each node) to atransparent one. There are several design choices in the deployment oftransparent optical networks.
First, in WDM networks a lightpath is required touse the same wavelength (this is referred to as a wavelength continuityconstraint in the literature) across the entire path unless intermediate nodesare equipped with wavelength convertors. Currently, commercially availablewavelength convertor technology uses OEO conversion. Thus, currentlytransparent optical networks must explicitly address the wavelength continuityconstraint. Alternatively, wavelength convertors are placed selectively at someof the nodes in the network (such an architecture where OEO conversion onlytakes place at a subset of nodes is called translucent) allowing the wavelengthcontinuity constraint to be relaxed somewhat. One advantage of networks withwavelength convertor is some modest improvement in network throughput.
Currently, there is considerable focus on the development of optical wavelengthconverters. All optical devices allow for the design of optical networks withlower power consumption (which is a very important area of research). Thus,some advantages of a transparent optical architecture with optical wavelengthconvertors are (i) significantly lower power consumption, and (ii) increasednetwork throughput.
Our paper focuses on transparent optical networks andassumes the availability of optical wavelength convertors. Thus the wavelengthcontinuity constraint can be relaxed.Inoptical communication, WDM is a technology that carries a variety of opticalcarrier signals on one fibre by using completely different wavelengths of laserlight.
This enables bifacial communication over one customary fibre with ininflated capacity. As optical network supports huge bandwidth; WDM network splitthis into a variety of small bandwidths optical channel. It permits multipledata stream to be transferred along a same fibre at the same time. A WDM systemuses a variety of multiplexers at the transmitter end, which multiplexes overone optical signal onto one fibre and demultiplexers at the receiver to separatethem apart. Usually the transmitter consists of an optical laser and modulator.The light supply generates an optical carrier signal at either fixed or atuneable wavelength. The receiver consists of photodiode detector whichconverts an optical signal to electrical signal. This new technology allowsengineers to increase the capacity of network without laying more fibre.
It hasmore security compared to other types of communication from tapping and alsoimmune to crosstalk. In this digital era the communication demand has increasedfrom previous eras due to introduction of new communication techniques. As wecan see there is increase in clients day by day, so we need huge bandwidth andhigh speed networks to deliver good quality of service to clients. Fiber opticscommunication is one of the major communication systems in modern era, whichmeets up the above challenges. This utilizes different types of multiplexingtechniques to maintain good quality of service without traffic, less complicatedinstruments with good utilization of available resources .
Wavelength DivisionMultiplexing (WDM) is one of them with good efficiency. It is based on dynamiclight-path allocation. Here we have got to take into consideration the physicaltopology of the WDM network and the traffic. We have taken performance analysisas parameter to analyse which type of topology is best suited to implement inreal life application without degrading quality of service (QoS). Wavelength Channel Multiplexing (WDM) is important technology used in today’s telecommunication systems. It has better features than other types of communication with client satisfaction. It has several benefits that make famous among clients such as WDM networks supports data to be transmitted at different bit rates. It alsosupports a number of protocols.
So there is not much constraint in how we want to send the data. So it can be used for various very high speed data transmission applications. WDM networks allows for wavelength routing. So in different fibre links the same wavelength can be used again and again. This allows for wavelength reuse which in turn helps in increasing capacity. WDM networks are also very flexible in nature.
As per requirement we can make changes to the network. Extra processing units can be added to both transmitter and receiver ends. By this infrastructure can redevelop to serve more number of people. WDM networks are extremely reliable and secure. Here chance of trapping the data and crosstalk is very low.
It also can recover from network failure in a very efficient manner. There is provision for rerouting a path between a source-destination node pair. So in case of link failure we will not lose any data.
The optical network has huge bandwidth and capacity can be as high as 1000 times the entire RF spectrum. But this is not the case due to attenuation ofsignals, which is a function of its wavelength and someother fibre limitation factor like imperfection and refractive indexfluctuation. So 1300nm(0.32dB/km)-1550nm (0.2dB/km)window with low attenuation is generally used. According to different wavelength there are threeexisting types namely- Wavelength Channel Multiplexing (WDM), Coarse WavelengthDivision Multiplexing (CWDM), Dense Wavelength Division Multiplexing (DWDM).The paper2says designed the Mat Plan WDM software for topology design, multi hour analysis &performance analysis. It isa MATLAB-based publicly available network planning tool for Wavelength-Routing (WR) optical networks, and it was fully developed by our research group.
His paper describes the multi-hour planning analysis extension included into the Mat Plan WDM version 3.It’s novel functionality allows the user to test planning algorithms which react under changes in the traffic demands. Multi-hour traffic patterns appear typically inbackbone WR networks that span over large geographical areas, where network nodes are situated in different time zones. A case study example is included to illustrate the merits of the tool. The articles 16 discuss the routing and wavelength assignment (RWA) problem in optical networks employing wavelength division multiplexing (WDM) technology.
Two variants of the problem are studied: static RWA, whereby the traffic requirements are known in advance, and dynamic RWA in which connection requests arrive in some random fashion. Both point- to-point and multicast traffic demands are considered. Input datafor communication network design/optimization problems involving multi-hour or uncertain traffic can consist ofa largest oftraffic matrices 17. These matrices are explicitlyconsidered in problem formulations for link dimensioning.
However, many of these matrices are usually dominated by others so only a relatively small subset of matrices would be sufficient to obtain proper link capacity reservations, supporting alloriginal traffic matrices. Thus, elimination ofthe dominated matrices leads to substantially smaller optimizationproblems, making them treatable by contemporary solvers. In their paper theydiscussed the issues behind detecting domination of one traffic matrix over another. They consider two basic cases of domination: (i) total domination when the same traffic routing must be used for both matrices, and (ii) ordinary domination when traffic dependent routing can be used. The paper isbased onour original results and generalizes the domination results known for fully connected networks. Due to power considerations 18, it is possible that not all wavelengths available in a fiber can be used at a given time.
In his paper, ananalytical model is proposed to evaluate the blocking performance of wavelength-routed optical networks with and without wavelengthconversion where the usable wavelengths in a fiber is limited to a certain maximum number, referred to as wavelength usage constraint. The effect of the wavelength usage constraint is studied on ring and mesh-torus networks. It is shown that the analytical model closely approximates the simulation results. It is observed that increasing the total number of wavelengths in a fiber is an attractive alternative to wavelength conversion when the number of usable wavelengths in a fiber is maintained the same. The paper 19 says while optical-transmission techniques have been researched for quite some time, optical “networking” studies have been conducted only over the past dozen years or so. The field has matured enormously over this time: many papers and Ph.D. dissertations have been produced, a number of prototypes and test beds have been built, several books have been written, a large number of start-ups have been formed, and optical WDM technology is being deployed in the marketplace at a very rapid rate.
The objective of this paper isto summarize the basic optical-networking approaches, briefly report on the WDM deployment strategies oftwo major U.S. carriers, and outline the current research and development trends on WDM optical networks.In WDM technology to be deployed werequire a physical topology. After topology designing we require routing andwavelength task to make it fully functional. Here we have taken three drawbackstatements as: To design 60 Gbps capacity topology and compare performancematrices and, to design 100 Gbps capacity topology and compare the performancematrices with 60 Gbps topology. A connection needs to be setup in the opticallayer so as to carry the data between the clients of the network.
The opticalconnection that is maintained between a source node, s and destination node, dis known as an optical path or light path. The difficulty of finding a path fora light path and assigning a wavelength to the light path is referred to as therouting and wavelength assignment problem (RWA). The problem of RWA is dividedinto two parts namely routing and wavelength assignment. We designed fourdifferent mesh network topologies (fully connected) having 6, 9, 12 and 15nodes. Also we have further added to design two9-nodes networks to analyse the link failure case. We designed an .
xml code todesign each network. The .xml contains the list of nodes and fibre links in thenetwork. Per node information is composed by the X and Y coordinates of thenode measured in kilometres over a Euclidean plane, number of E/O transmitters,O/E receivers, node population, node type (or node level), number of nodes andthe name of each node. Per link information is the maximum number ofwavelengths per link and the number of optical fibres. MatPlan WDM is a goodsimulation tool to analyse different network topologies and performancematrices.
With this we can use different types of designing algorithm like MILP(Mixed integer Linear Programming). For this we need TOMLAB (requireregistration) which allows many more Functions like wavelengthconversion/without wavelength conversion, with traffic losses/without trafficlosses, losses cost per Gbps, cost per electronically switched Gbps, maximumlight path distances. Its GUI (Graphic User Interface) allowsthe user to carry out full multi-hour test for a prebuilt or user definedmulti-hour planning algorithm and virtual topology analysis. We can worktowards dynamic Analysis which allows GUI to test online optimizationalgorithms to react to high level traffic connections arrivals, terminationsand to do high level of performance analysis with good accuracy.In such an optical WDM network architecture, thefailure of a network component such as a fiber can lead to the failure of allof the light paths that traverse the failed fiber. Since each light path isexpected to operate at a rate of few Gb/s, a fiber failure can lead to asignificant loss of bandwidth and revenue.
For an IP-over-WDM network, twomethods for providing protection are as follows: 1) provide protection at theWDM layer (i.e., set up a backup light path for every primary light path), or2) provide restoration at the IP layer (i.e., overprovision the network sothat, after a fiber cut, the network should still be able to carry the sameamount of traffic as it was carrying before the fiber cut). Thetwo problems are how to interconnect WDM ring networks as well as how to groomthe traffic in interconnected rings.Herewe have used MatPlanWDM0.61 as a stimulation tool to simulate our topologies.
It takes physical topologies and traffic data for different network topologies.Here performance analysis of the four topologies has been done using MatPlanWDM0.61simulator. We have to give topologies in .xml and traffic file in .traffformat.
The algorithm we used here is shortest path algorithm. After that wehave selected sweep parameters with lower and upper limits and number of sweeppoints to start simulation. We have then compared the results of the four 60gbps network topologies and analysed them based on different parameters toconclude which network topology suits best to provide best quality of service. Number of light path increases with increase in traffic demand as light paths are created as per demand. It decreases with increase inWCC, because as the capacity for each channel increases, the number oflight path will decrease to maintain the total offered traffic. More number of light paths are desirable for better routing. Therefore network having more number of nodes is preferable.
For a topology to be implemented in reallife application it has to have minimum delay, low network congestion rate,maximum number of possible light paths and high Single Hop Traffic/OfferedTraffic. In case of normal assumption one can think that delay will increasewith increase in number of nodes. It depends upon number of light paths. Sowith increase in number of nodes, the number of light paths increases fromsource node to destination node in shortest path algorithm. So the queuingdelay decreases, decreasing the overall delay.
Since we can get more lightpaths, so delay decreases with increase in nodes. We can see the networkcongestion, number of light paths, single hop traffic/offered traffic increaseswith increase in number of nodes.Recent advances in the field of opticalcommunication have opened the way for the practical implementation of WDMnetworks. After going through several papers we have found out that fordetermining Quality of Service the effect of network architecture is not takeninto account.
So we have studied and planned to design four different networksand simulate them with different scenario to determine the performancematrices, which are called QoS (Quality of Service) parameters. In this work,we have decided of using the simulation tool MatPlanWDM0.61, to study WDMnetworks and their performance analysis, which is freely available. It is anexcellent framework for designing & development of topology with variousfeatures.
Studying different research papers we have concluded that if there isless number of nodes with high capacity, then delay will be more. If number ofnodes is more as well as high capacity then network congestion will be more. Sowe have to choose a minimized output to maintain a better QoS.
