Proc. of Int. Conf. on Emerging Trends in Engineering and Technology Performance Analysis on Various Wavelength Assignment Algorithms with Traffic Grooming Vikas Kaushik 1, R.S Chauhan 2 1 JMIT Radaur/ECE Department, Radaur, India Email: er.vikas7@gmail.com 2 JMIT Radaur/ECE Department, Radaur, India Email: chauhan@jmit.ac.in Abstract Wavelength Division Multiplexed switching networks are important for the future transport networks. This paper analyzes the performance of various wavelength assignment algorithm and their effect on the blocking probability of the connection request in the optical network with traffic grooming. The wavelength assignment is a unique feature in wavelength routed network that distinguish them from the conventional networks. The wavelength assignment algorithms are classified as First fit, Random fit, Most used, Least used and wavelength conversion algorithms. Simulation is done for the 16 node optical ring network. The experimentation results indicate that the most used algorithm achieves reduced network blocking rate with and without traffic grooming. Index Terms WDM Networks, Blocking Probability, First-Fit method, Random-Fit method, most used, least used, least loaded. I. INTRODUCTION To fulfill the demand of bandwidth required for the services such as transmission of audio, video and data using video on demand etc, the WDM optical network plays an important role in optical networks. Thus, with the help of WDM technology, it is possible to transmit traffic on different wavelength within the same optical fiber simultaneously. But in early days, the optical fiber communication was only confined to do transmission of single fiber channel, but now a days optical network satisfies the requirement of high transmission. Since the data rate increases at rapid rate so in order to meet the efficient utilization of the fiber capacity WDM(wavelength division multiplexing) network came into scenario to meet the demands of high bit rate. WDM enables the efficient utilization of optical fiber by dividing its tremendous bandwidth into a set of disjoint wavelength bands, which are referred as wavelengths. Each of these wavelengths supports one communication channel which corresponds to an end user operating at an arbitrary speed and helping in overcoming the optoelectronic mismatch between multiple terabit per second bandwidth of optical fiber and gigabit per second electronic processing speeds at end user. WDM based optical network enables the fiber to provide the large throughput as each of these multiple channels have assigned different wavelengths and transmitted over the same fiber simultaneously. Wavelength division multiplexing also solves the problem of dispersion which came into the effect during transmission in the fiber by keeping the transmission rates of each channel at reasonably low levels and achieving a high data rate by combining many channels [17]. Based on different wavelength patterns the WDM networks are categorized as conventional or coarse WDM and dense WDM. Conventional WDM system provides up to 16 channels whereas dense WDM provides denser channel spacing on same transmission window. The performance of WDM bas optical network primarily depends on routing and wavelength assignment techniques used. These techniques are evaluated in DOI: 03.AETS.2013.3.139 Association of Computer Electronics and Electrical Engineers, 2013
terms of call/connection blocking probability in network. The connection blocking is the probability that an incoming connection or call request is blocked or denied, due to insufficient resources between source and destination. For every dynamic connection request, a light path is needed to be established, otherwise connection is blocked. A network s performance is inversely proportional to the amount of connection blocking in the network. The factors which causes the connection blocking in network are: insufficient network resources like wavelength or bandwidth, lack of wavelength convertors in network, routing and wavelength assignment decision made on outdated network state information. Insufficient network resources mean lack of available wavelength and convertors. The wavelength continuity constraint increases the blocking probability [17]. According to this constraint the same wavelength should be maintained on all links from source to destination. This constraint affects the performance of the network by the blocking connection requests because of unavailability of common wavelength on intermediate links. The RWA problem on sub wavelength demands with the objectives of minimizing the network cost and optimizing network throughput is called traffic grooming. The traffic grooming concept came in 1990 s to address the gap in between channel capacity and individual traffic demands in optical networks. The problem of multiplexing and routing low speed traffic requests over light path as well as determining their wavelength assignment is known as traffic grooming technique[10]. Bandwidth requirement of end users is too less as compare to big companies the end users are in Mbps, that is too less as compare to capacity of an entire wavelength. So to accommodate low speed streams and to reduce wastage of bandwidth, traffic grooming technique is used. According to their functionalities traffic grooming is of two types which are single hop and multi hop grooming technique. Wavelength continuity constraint can be relaxed by using wavelength convertors. A wavelength convertor is a single input/output device that converts the wavelength of an optical signal arriving at its input port to different wavelength as the signal departs from its output port, but otherwise leaves the optical signal unchanged. Different levels of wavelength conversion shown in figure 1. Figure 1: Levels of wavelength conversion Figure 2: Block diagram for wavelength assignment algorithms. 289
Wavelength conversion increases the routing choices for a given light path, resulting in better performance in terms of less blocking probability. It also reduces the bandwidth loss that results in better bandwidth utilization. However such convertors are expensive. The routing and wavelength assignment (RWA) can b split up in to two parts: first choosing the route then assigning the wavelength/channel to that route. In routing process the routes are chosen based on shortest path selected. Assignment of wavelengths to the connection request is depicted in figure 2[15]. II. VARIOUS WAVELENGTH ASSIGNMENT ALGORITHMS ARE AS FOLLOWS: A. First fit (FF) wavelength algorithm In this method first the free wavelengths of traffic matrix are sorted in non decreasing order. FF strategy always chooses the lowest indexed wavelength from the list of free wavelengths and assigns it to the connection request. When the request is completed the wavelength is added back to the free wavelength set. B. Random fit (RF) wavelength algorithm In this method, a set of free wavelengths on particular path is determined. RF algorithm determines which wavelengths are available and then choose the wavelength randomly amongst the available set of free wavelengths. C. Most used (MU) and least used (LU) wavelength algorithm In most used algorithm, whenever the connection request is made it get to be allocated by the wavelength which is using on the greatest number of fibers in the network. Least used (LU) wavelength assignment is similar to the most used algorithm except in LU algorithm the least used wavelength in the wavelength is assigned. Least Loaded wavelength assignment selects the least loaded wavelength on a most loaded link. This wavelength available on maximum number of fiber for transmission. III. LITERATURE SURVEY Barry and Pierre [1] introduced a traffic model for circuit-switched all-optical networks which they then use to calculate the blocking probability along a path for networks with and without wavelength changers. A,Birman et al. [2] examined that the wavelength converters devices are much expensive so much research work focuses on sparse wavelength conversion, in which only a part of network nodes have the capability of wavelength conversion whereas in full wavelength conversion all the nodes of a optical networks having such wavelength conversion capability. The blocking probability for different optical network models under fixed routing and least loaded routing is analyzed using generalized reduced load approximation techniques. Mokhtar and Azizo glu [3] considered routing and wavelength assignment in wavelength-routed all-optical networks supporting circuit switching technique. The approach to address this issue consider the two aspects of the problem disjointly by first finding a route from a predetermined set of candidate paths and then searching for an appropriate wavelength assignment. The analysis can also accommodate networks with multiple fibers per link. The blocking performance of the proposed adaptive routing algorithms is compared along with their computational complexity. Kai-Yeung Siu et al. [4] analyzed the throughput performance of online algorithm for the problem of routing and wavelength assignment in single hub WDM ring network. Mewanou and Pierre [5] proposed two new heuristic algorithms to manage optical routing based on link-state and reduce blocking probability of request arriving in the network. The technique used is similar to fixed paths least congested (FPLC) routing by analyzing the first k link on each path whereas the second is based on an estimation of the link-congestion in the proposed network. These both techniques achieve good performance. Paramjeet singh et al. [6] examined RWA problem on wavelength division multiplexing networks without wavelength conversion. He et al. [7] discussed and presented a new heuristic offline wavelength ordering algorithm for wavelength allocation. Also they studied the impact of guaranteeing QOS, by combining both BER and latency constraints, on the performance of wavelength assignment algorithms in shortest path (sp) and fixed alternate routing. From the results it is observed that heuristic algorithm minimizes the crosstalk due to adjacent wavelength power leaking through the WDM demultiplexers 290
Sharma et al. [8] proposed that Wavelength Division Multiplexed switching networks are considered as an important candidate for the future transport networks. Analysis is based on the calculation of Blocking Probability in the network. Analysis shows that Blocking Probability is less when wavelength conversion is performed as compared to when no wavelength conversion takes place. Further, in case of no wavelength conversion, First-Fit method gives less blocking probability as compared to Random-Fit method. He et al. [9] proposed that the quality of an optical signal degrades due to physical layer impairments as it propagates down the length. As a result, the signal quality at the receiver of a lightpath may not be sufficiently high, leading to increased call blocking. The authors derived the analytical expressions for the total blocking probability for first-fit wavelength assignment for networks suffering from transmission impairments. The analysis is also applicable to first-fit algorithms with different static channel orderings. Farahmand et al. [10] focused on traffic grooming WDM optical network with dynamic traffic patterns. Through extensive simulation authors shows tha the proposed approaches lead to lower request blocking probability. Amit Wason et al. [11] suggested a model for WDM optical networks and also proposed a algorithm based on most used wavelength assignment for minimizing the blocking probability. These results are then compared with the conventional wavelength assignment algorithms such as first fit, random fit and most used wavelength assignment algorithms. Amit wason et al. [12] given a mathematical model for reducing the blocking probability of the WDM optical network. Their proposed model has low implementation complexity and its computation is quite efficient. This model is also used to evaluate the blocking performance of NSFNet topology and used to improve its performance. Sharma et al. [13] introduced a Fiber Optic Simulator (FOSP), used to optimize Physical properties for achieving high speed. The wavelength assignment algorithms (First Fit, Random Fit) & WDM networks with wavelength conversion is analyzed in terms of their blocking probability. It was analyzed that WDM networks with wavelength converters gives the best performance among all the wavelength assignment techniques used. Wang and Wen [14] proposes two lightpath-level active rerouting algorithms, which are the least resources rerouting algorithms and the load balanced rerouting algorithms. Simulation results show that the proposed load-balanced active rerouting algorithm yields much lower connection blocking probability than the least resources rerouting algorithms. This result shows that the performance of the networks can be significantly improved. Singal and Kaler [15] proposed a new simulation technique for the performance analysis of wavelength assignment algorithms such as First-fit, Random, Most-used with Wavelength conversion Algorithms in WDM unidirectional optical ring network. The blocking probability of various algorithms with the variation in number of events has been compared. The network performance under the wavelength conversion algorithms is best but there is burden of using expensive hardware. Without the need of wavelength converter the Most-used algorithm performs better than the Random and First-fit algorithms. Arunachalam et al. [16] analyzed the performance of wavelength assignment strategies to reduce the network block rate of request. Simulation results indicates that the round robin assignment algorithm achieves reduced network block rate with increased throughput. IV. SIMULATION The experiment is carried out to analyze the performance of various wavelength assignment algorithms. The parameters used for simulation are variable traffic and fixed wavelength. The ring topology with sixteen nodes is shown in figure 3. Figure 3: Optical ring network with 16 nodes 291
The simulation is done on the optical ring network model which is having 16 nodes with the fixed wavelength 7 and the variable offered traffic on each of the link between the nodes. The nodes used in the model are of non grooming type. The link capacities matrix which is of [16][16] in the case of ring network and the arrival rate matrix of order [16][16] is also used in the proposed model. V. RESULTS The blocking probability of network depends on length of the route, free wavelengths, and number of channels. We have fixed the wavelength as 7 and traffic load varies from 0 to 4. Here the total traffic blocked, average blocking probability and average of average blocking probability of algorithms such as first fit, random fit, most used, least used, least loaded and most loaded is compared with each other as shown in figure 4 to figure 6 without traffic grooming on the other hand with traffic grooming are shown in figure 7 to figure 9. Figure 4: Total Blocked. Figure 5: Average Blocking Probability. Figure 6: Average of average Blocking Probability. 292
Figure 7: Total blocked Figure 8: Average Blocking Probability. Figure 9: Average of Average Blocking Probability. VI. CONCLUSION The blocking probability in an optical network depends on the Network hierarchy as well as offered Traffic to that network. The wavelength assignment problems are the major problems of WDM networks. Hegons simulator is used for performance analysis of optical network using first fit, random fit, most used, least used, least loaded and most loaded wavelength assignment algorithms. The simulation results obtained for calculating the blocking probability for each of the algorithms has been compared. The most used algorithm performs better than all other algorithms without the need of wavelength converters. The blocking performance of WDM network was analyzed for optical ring network having 16 nodes with fixed wavelength. Simulation results obtained for calculating the blocking probability, total blocking probability, average blocking probability, average of average blocking probability and overall blocking probability for each of the algorithm are compared and it is analyzed that the most used offers least average of average blocking probability with and without traffic grooming which are 0.203639 and 0.211229 respectively. 293
REFERENCES [1] Barry A.Richard, Humblet A. Pierre, Models of Blocking Probability in All- Optical Networks with and without Wavelength Changers, IEEE, Volume 14, Issue 5, pp:858 867, 1995. [2] A.Birman, computation approximation blocking probabilities for a class of all optical network, IEEE J.select.Areas commun.14(5)(1996)852-857. [3] Mokhtar, Azizoglu, Adaptive Wavelength Routing in All-Optical Networks, IEEE TRANSACTION ON NETWORKING, VOL. 6,1998 [4] Kai-Yeung siu, Ching Law, online routing and wavelength assignment in single hub WDM rings, IEEE, PP. 2111-2123, vol. 18, October-2000. [5] Rolland MEWANOU, Samuel PIERRE, Dynamic routing algorithms in all optical networks IEEE CONFERENCE, Vol.[2],Page(s): (773-776),2003. [6] Paramjeet singh, Ajay K. sharma, Shaveta Rani, Routing and Wavelength assignment strategies in optical networks, optical fiber technology, vol. PP. 191-197, 2007. [7] Jun He, maite Brandt-pearce and suresh subramaniam, QOS- Aware Wavelength With BER and Latency constraints For all optical Networks, Journal of Lightwave Technology, vol.27,no. 5, Page(s).462-474,2009 [8] Sharma Amandeep, Kaur Jaspreet, Sachdeva Sandeep, Data Transfer Efficiency Enhancement & Effect of Blocking Probability Research- 2KX, HEC,, Jagadhri, pp. 526-529.2010 [9] Jun He, maite Brandt-pearce and suresh subramaniam, Analysis of Blocking Probability for First-Fit Wavelength Assignment in Transmission-Impaired Optical Networks, OPT. COMMUN. NETW, VOL. 3, NO. 5, Page(s): 411-425,2011 [10] Farid Farahmand, Xiaodong huang and Jason p.jue Efficient traffic grooming algorithm in wdm mesh networks with drop and continue architecture,2008 [11] Amit wason, RS. Kaler, Wavelength assignment algorithms for WDM optical networks optic 122(2011) 877-880. [12] Amit Wason, RS Kaler, Blocking probabilities calculation in wavelength routed all optical networks, optic 122 2011. [13] Sharma Amandeep, kaur jaspreet, Sachdeva Sandeep, Optimized WDM Network with Consideration of Lesser Blocking Probability & Shortest Path Selection, International conference on signal processing computing and control, Page(s): 1-6,2012 [14] Wei Wang, Yen Wen, Lightpath-Level Active Rerouting Algorithms in All-Optical WDM Networks with Alternate Routing and Traffic Grooming, Inernational conference on information networking, Page(s): 42-46.2012 [15] Singal.A and Kaler R.S, Performance Evaluation of Algorithms for Wavelength Assignment in Optical Ring Network, Second International Conference on Advanced Computing & Communication Technologies,2012 [16] M. Arunachalam, S. Nandhini, P. Pavithra, Experimental analysis of wavelength assignment strategies in optical N/W, International journal of emerging technology and advanced engineering, Volume 2, issue 8,August-2012. [17] V. Kaushik, RS. Chauhan, A review on various wavelength assignment algorithms for WDM network Yamuna Journal Of Technology and business Research, volume-3, (issue1-2),2013. 294