Dynamic RWA in WDM optical network

As optical wdm networks are becoming the backbone building blocks of twenty first century’s networks due to their capacity of carrying huge bandwidth. So evaluating the performance of optical wdm networks and optimizing it, is need of the day. For this the problem of routing and wavelength assignment is of significance importance. Researchers have done extensive simulations for both static and dynamic routing and wavelength problem. In this paper we have presented various algorithms for solving the dynamic RWA problem in literature.Also we have prepared a optical wdm network using which RWA problem is analysed.


Introduction
Nowadays optical wdm networks are emerging technologies networks that can provide routing, grooming, and restoration at wavelength level [2].WDM technology of crucial importance for satisfying the ever increasing capacity requirements in telecommunication networks.Wavelength routed WDM networks can exploit the large bandwidth of optical fibers by dividing it among different wavelengths.These networks are equipped with configurable WDM nodes which enable us to set up and tear down all-optical channels, called light paths, between pairs of nodes.Given a network and a set of light path requests, the Routing and Wavelength Assignment (RWA) problem attempts to route each light path request, and to assign wavelengths to these routes subject to the following constraints [1,13] .The RWA problem can be categorized into two types, static RWA (SRWA) and dynamic RWA (DRWA).In a wavelength-routed optical network (WRON), the traffic demands can be either static or dynamic In static RWA, all the connection requests are known in advance and the objective of RWA is to assign routes and wavelengths to all the light paths globally while keeping the amount of network resource required minimal [5].In a dynamic traffic pattern, a light path is set up for each connection request as it arrives and it is released after some finite amount of time.One of the challenges involved in designing wavelength routed networks with dynamic traffic demands is to develop efficient algorithms and protocols for light paths establishing [7, 8, 10 and 11].In all-optical WDM networks, each of the connections going through a link is assigned a wavelength.If a connection is assigned the same wavelength on all links along its routing path, the signal is able to travel from the source node to the destination node using the same wavelength.In case that the connection has to be assigned two or more wavelengths on different links, one or more nodes along its routing path must have the capability of converting the signal from one wavelength to another wavelength.If all nodes along the routing path of a connection are incapable of converting a wavelength to another wavelength, the connection must be assigned the same wavelength on all links along its path.Otherwise, the connection is blocked.This is known as the wavelength continuity constraint.A wavelength which is available on all links along a routing path is referred to as a common available wavelength along the routing path [6].In addition, light paths that share a common physical link cannot be assigned the same wavelength.This is called the wavelength clash constraint [13,1].There are number of algorithms available in literature for dynamic routing and wavelength assignment problem and rsearchers too have used optimization techniques and have formulated various new algorithms to optimize the performance of optical wdm network in terms of dynamic wavelength and assignment problem.In this paper we have presented the review of various techniques available for solving dynamic routing and wavelength problem.

Fixed Path Routing
It uses only one route from source end to destination .In this approach the network resources are not properly utilized.[ ICAET 2016 - It uses alternate paths from source to a destination .If the first is unavailable then second path is used.[15] In dynamic routing in DRW1 algorithm no weight assignment is done all links have same weight.First all possible paths between the source and destination is calculated then request is routed over the shortest path if wavelength is available else alternate shortest path is used.In DRW2 less weight is assigned to nodes which are used less.As the request arrived it is routed to the nodes having less weight if wavelength is available if not then request is routed through next higher weight nodes and the process is repeated.In DRW3 algorithm less weight is assigned to the nodes which are less used and are nearest and then routing of request is done through routes having minimum weight if wavelength is available else next higher weight is tried and the process is repeated [9] Another algorithm is proposed in literature which first selects K edge disjoint least weighted paths according to the current network state and then one such path is considered to route the request taking into account not only the load ,length of the path but also number of transceivers in the nodes along the path.afterthis an auxiliary graph is generated and request is routed on least weight route on the graph and wavelength is assigned for the same.[14]

Fixed alternate shortest Path routing with least Priority wavelength assignment Strategies
In this algorithm the wavelength information table is made according to the use of the each wavelength in the network then the lightpath is set on shortest or alternate shortest path with the least used wavelength if available else the route call is blocked.[16]

Exhaust Algorithm
It searches all the possible routes between source and destination and then finally selects the best route .So utilizing this route provides the flexibility of selecting any route according to the wavelength available.Source Initiated Reservation Algorithm In this algorithm the network resources are utilized by reducing control overhead .In this the route table is generated first and then resource and wavelength availability is updated .Next the route table is arranged according to minimum hops and maximum wavelengths available and route is searched according to that.If wavelength allotted the request is processed else the request is blocked and then whole process is repeated.[15]

Traffic Intensity Based Fixed -Alternate Routing
In this algorithm the traffic is splitted for each source destination pair onto the disjoint routing paths by solving the nonlinear multicommodity flow optimization problem.Further the routing path between end users is sorted in decending order according to the amount of traffics of source destination pair assigned to the routing paths.For the request arrived this algorithm tries the multiple routing paths between the particular source destination pair one by one in the sorted order to establish a connection until the request is processed successfully otherwise the request is blocked .Results proved that this algorithm have lower blocking probability as compared if sorting is done according to the hop count instead of traffic intensities.[3]

Priority and Maximum Revenue based Routing
In this algorithm two types of requests are kept in two light path request matrixes namely high priority matrix (HPM) and low priority matrix (LPM).High priority matix requests are entertained first before LPM requests.
According to this algorithm the light paths in HPM must be fulfilled and revenue for the same is fixed while the light paths in the LPM can be set up selectively to optimize revenue when there is not enough uresource to accommodate all.[

Simulation Results
For analyzing the problem of routing and wavelength assignment we have prepared a wdm network in which each link is bidirectional and taken eight wavelengths per link.For routing shortest path algorithm is used and first fit for wavelength assignment.The traffic considered is dynamic in nature which follows the poisson process.We have analysed and calculated the thoughput for two types of applications i.e FTP and email .The simulation is carried on opnet modeler.

Conclusion
The traffic is dynamic in nature so the load is varying w.r.t time.Throughput is the performance metric taken in this paper which is observed lesser compared to the load offered to the network which indicates the blockage of requests offered to the network.The throughput and blocking of the network can be optimized with some nature inspired technique in future.

Figure 1 .
Figure 1.Load versus time for email application.

Figure 3 .
Figure 3. Throghput versus time and load for email.

Table 1 .
Data sheet for load and throughput.