An Evolutionary Algorithm Approach for Dedicated Path Protection Problem in Elastic Optical Networks

Size: px
Start display at page:

Download "An Evolutionary Algorithm Approach for Dedicated Path Protection Problem in Elastic Optical Networks"

Transcription

1 An Evolutionary Algorithm Approach for Dedicated Path Protection Problem in Elastic Optical Networks Mirosław Klinkowski Department of Transmission and Optical Technologies, National Institute of Telecommunications, Warsaw, Poland Elastic Optical Network (EON) architectures are a very promising solution for next generation optical networks. In EON, a basic problem in network design and operation is the problem of Routing and Spectrum Allocation (RSA). In this work, we focus on RSA in a survivable EON with Dedicated Path Protection (DPP) consideration. Since RSA is a difficult problem itself, we develop an evolutionary algorithm (EA) with the aim to support the search for optimal solutions. We investigate the effectiveness of algorithm for a set of survivable network scenarios. Evaluation results show that the proposed algorithm outperforms reference algorithms from the literature. 1 KEYWORDS: elastic optical networks, evolutionary algorithm, network optimization, dedicated path protection, routing and spectrum allocation INTRODUCTION The communication networking is undergoing fundamental changes which are governed by the development and integration of diverse network technologies and applications within a unified Internet Protocol (IP) packet-oriented network infrastructure. The observed trends result in an exponential growth of the traffic generated by network users and in the need for efficient and flexible optical transport networks able to support multi-rate and huge-bandwidth (100+ Gb/s) connections. The currently deployed wavelength division multiplexing (WDM) optical networks operate within a rigid frequency grid and with single-line-rate transponders making use of single carrier modulation 1 This work has been supported by the Polish National Science Centre under grant agreement DEC- 2011/01/D/ST7/05884.

2 M. Klinkowski techniques. The evolution path of optical communication networks leads toward elastic optical networks (EONs) in which advanced single-carrier modulation formats (such as Phase-Shift Keying and Quadrature Amplitude Modulation) and multi-carrier modulation techniques (such as Optical Orthogonal Frequency Division Multiplexing) are applied for mixed-line-rate transmission and where elastic access to spectral resources is achieved within flexible frequency grids (flexgrid) (Gerstel et al. 2012; Wei et al. 2012). Future optical networks will utilize spectrum resources in optical fibres more efficiently, according to the transmission path characteristics and bandwidth requirements. We refer to some recent papers for more details on EON architectures (Jinno et al. 2009; Meloni et al. 2011) and for reports on proof-of-concept EON experiments (Geisler et al. 2011; Cugini et al. 2012). ITU-T has recently revised the G recommendation and included the definition of a flexible WDM grid (ITU-T G.694.1). According to G.694.1, the frequency spectrum in an optical fiber link is divided into narrow frequency segments (we refer to them as slices). The optical path (lightpath) is determined by its routing path and a channel, which consists of a flexibly (adhoc) assigned subset of slices around a nominal central frequency. Elastic spectrum allocation in EONs differs with channel assignment in fixed-grid WDM networks in that the channel width is not rigidly defined but it can be tailored to the actual width of the transmitted signal. Due to this difference, ordinary Routing and Wavelength Assignment (RWA) algorithms are not appropriate for EONs. In EONs, the problem of finding unoccupied spectrum resources so that to establish a lightpath is called the Routing and Spectrum Allocation (RSA) problem. RSA concerns assigning a contiguous fraction of frequency spectrum to a connection request - we refer to it as the spectrum contiguity constraint - subject to the constraint of no frequency overlapping in network links. The RSA optimization problem is NP-hard (Klinkowski and Walkowiak 2011) and it is more challenging than RWA in fixed grid networks due to the existence of the spectrum contiguity constraint. Offline RSA has been addressed with both Integer Linear Programming (ILP) (e.g., see Christodoulopoulos et al. 2011; Klinkowski et al. 2011; Velasco et al. 2012), metaheuristics (Christodoulopoulos et al. 2011, Klinkowski 2012, Gong et al. 2012), and heuristic algorithms, which are based either on ILP problem 2

3 Evolutionary Algorithm for DPP in EON relaxations (Christodoulopoulos et al. 2011) or on greedy, sequential processing of demands (Christodoulopoulos et al. 2011; Klinkowski et al. 2011). So far, there have been proposed several solutions for survivable EONs in the literature. Regarding offline network design, survivable RSA algorithms for dedicated path protection (DPP) and shared backup path protection (SBPP) in a ring network have been studied in Takagi et al DPP in a network with generalized connectivity has been addressed with both ILP formulation (Klinkowski and Walkowiak 2012) and heuristic algorithms (Patel et al. 2011; Klinkowski and Walkowiak 2012). Concurrently, an ILP formulation (Eira et al. 2012), a metaheuristic algorithm (Castro et al. 2012), and heuristic algorithms (Eira et al. 2012; Liu et al. 2012) have been proposed for SBPP. Eventually, in Assis et al. 2012, a MILP formulation and an heuristic algorithm have been presented for the so-called squeezed protection scheme in which the protection of a lightpath may not cover its entire bandwidth but it is allowed to be partial. All these solutions assure single-link failure survivability. Evolutionary algorithms (EA) are an effective population-based mataheuristic approach for providing near-optimal solutions for large-scale optimization problems (Talbi 2009). In this paper, we employ EA in the search for optimal RSA solutions in an EON with dedicated path protection (DPP) consideration and static traffic demands. In particular, we develop an efficient EA-based algorithm which, as numerical results show, outperforms other reference algorithms for the offline RSA/DPP problem in EON. This paper extends our recent work presented in a conference paper (Klinkowski 2012). Initially, we have focused on a DPP scenario in which a pair of working and backup lightpaths has allocated the Same Channel (SC), i.e., the same segment of spectrum. Here, we contribute with an extension and evaluation of the EA algorithm adapted to an alternative DPP scenario, namely, with Different Channel (DC) allocation. It should be mentioned that, however, an evolutionary approach for optimizing RSA has been recently presented in Gong et al. 2012, still, to the best of our knowledge, the application of EA in survivable EON scenarios has not been studied in the literature prior to Klinkowski 2012 and this work. Besides, Gong et al propose a twopopulation based algorithm which differs with our one-population based EA approach. 3

4 M. Klinkowski The article is organized as follows. In the following section, we discuss the details of both SC and DC survivable network scenarios and formulate the RSA/DPP problem. In the next section, we describe our EA-based approach to solve RSA/DPP. Then we present numerical results. Finally, we conclude the work. PROBLEM FORMULATION In this section, we formulate an offline problem of Routing and Spectrum Allocation with Dedicated Path Protection in an EON with static traffic demands and subject to single-link failures. Among network survivability schemes, connection recovery through path protection, in which back up network resources (i.e., lightpaths) are provisioned in advance for each connection, is preferred due to its quick recovery time. In DPP, each connection has its own backup resources, on the contrary to SBPP, in which backup resources can be shared between the demands whose primary paths are not likely to fail at the same time. Similarly as in Patel et al and Klinkowski and Walkowiak 2012, we distinguish two DPP scenarios, namely: a) with Same Channel (SC) allocation and b) with Different Channel (DC) allocation; we denote the resulting problems as RSA/DPP/SC and RSA/DPP/DC, respectively. Case SC is a cost-effective scenario in which the transponders are shared between primary (working) and backup connections and a traffic demand has allocated the same segment of optical frequency spectrum (i.e., channel) on its primary and backup path. Such a solution reduces the network cost and alleviates the connection switching time, as discussed in Patel et al In the DC case, the SC constraint of having the same operating channel in both primary and backup connections is relaxed, by these means, reducing spectrum requirements in the network (Patel et al. 2011; Klinkowski and Walkowiak 2012). One way to implement this scenario is to install dedicated (for both working and backup lightpaths) transponders. However, this is an expensive approach since it doubles the cost of transponders installed in the network. An alternative solution is to use laser-tunable transponders which are able to modify the central frequency of the transmitted optical signal (Buus 2006). Since such technology is commercially available and adopted in WDM 4

5 Evolutionary Algorithm for DPP in EON core network applications (Blum 2010), we believe that it is also a viable solution for elastic optical networks. The considered RSA/DPP problem can be formally stated as: Given: 1. an EON represented by a graph G=(V,E), where V denotes the set of nodes, and E denotes the set of fiber links connecting two nodes in V; 2. a frequency spectrum, the same for each link in E (without loss of generality), with the flexgrid represented by an ordered set of frequency slices S={s 1,s 2,...,s S }; 3. a set D of static traffic demands to be transported with path protection guarantees; each demand d is represented by a tuple (o d, t d, n d ), where o d and t d are source and destination nodes respectively, and n d is the requested number of slices. Find a primary lightpath and a backup lightpath over the EON for every transported demand subject to the following constraints: 1. spectrum contiguity: for each demand, the slices should be allocated each next to the other (i.e., adjacent) in the frequency spectrum, 2. spectrum continuity: for each demand, the subset of allocated slices should be the same for each link on the selected routing path, 3. slice capacity: a slice in a link can be allocated to one demand at most, 4. same channel (applies only for RSA/DPP/SC): for each demand, the primary and backup lightpath should have allocated the same subset of slices, and with the objective to minimize the width of spectrum (denoted as Φ) required in the network. Here, we assume the same optimization objective as in Takagi et al. 2010, Patel et al. 2011, and Eira et al

6 M. Klinkowski The requested number of slices n d is a function of the requested bandwidth, the modulation technique applied, the slice width in the flexgrid, and the guard band introduced to separate two spectrum adjacent connections, among others. For a given network scenario, where both the transmission parameters and flexgrid definitions are given, there is a relation between n d and the requested bandwidth (e.g., see Sec. II in Christodoulopoulos et al. 2011). Therefore, without loss of generality, we express the demand volume by means of the number of slices. Assuming the flexgrid definition proposed in G and its requirement to have the spectrum allocated symmetrically around a central frequency, n d is considered to be an even number, i.e., n d {2i: i Z + }. For more details, we refer to the G recommendation. In this paper, we consider that a set of candidate routing paths is given. Let P denote the set of all candidate paths. Each path p P is identified with a subset p E. Let Q denote the set of pairs of paths (p,q), where p P is a primary path and q P is a backup path. Let Q d denote the set of path pairs for demand d D; each set Q d comprises only paths that have the origin in o d and the termination in t d. In this work, we assume each Q d has the same number of candidate path pairs, which is denoted as k (i.e., Q d =k). Apart from that, we assure the network is resilient to single link failures by assuming that Q is a set of link disjoint path pairs. Let N be the set of the numbers of requested slices, i.e., N={n: d D,n d =n}, and n max =max{n N} and n min =min{n N} be, respectively, the largest and the smallest number in this set. In the remainder of this paper, we consider that the set of demands D={d 1,...,d i,d i+1,...,d D } is an ordered set and, in particular, the demands are sorted in decreasing order of n d, i.e., i we have n di n d(i+1). Eventually, let D n be the set including all demands with n requested slices, i.e., D n ={d:d D,n d =n}, n N. EVOLUTIONARY ALGORITHM FOR RSA/DPP In this section, we present a EA metaheuristic-based algorithm for the off-line RSA/DPP problem in EON. In the following, we describe our approach to solve the problem and then we focus on EA implementation details. 6

7 Evolutionary Algorithm for DPP in EON Algorithm Approach To solve RSA/DPP, we propose a hybrid approach in which we combine EA with a greedy RSA algorithm. The role of EA is to look for a sequence of demands which, when processed by the RSA algorithm, minimizes the occupied spectrum width (Φ). In details, the RSA subroutine serves the demands, one-by-one, in an order determined in the EA phase. Then, having performed RSA, we calculate Φ by counting the number of slices s S that are occupied in network links:, (1) where x s =1 if slice s S is occupied in any link e E, and equal to 0 otherwise. The value of Φ indicates the quality of solution and, by this means, it supports EA in the search for better sequences of demands (the details are discussed in following subsections). RSA Subroutine As the greedy RSA algorithm we use the MSF (Most Subcarriers/Slices First) algorithm presented in Christodoulopoulos et al. 2011, which we adapt to the DPP context. MSF performs RSA for a set of demands that are arranged and processed in decreasing order of the number of requested slices. If we assume that s dp denotes the largest slice index among the slices allocated to demand d on routing path p, then MSF looks for such path p * (among all candidate paths) and such unallocated spectrum segment on this path (consisting of n d slices) that minimizes s dp. Then, demand d is allocated on path p * within spectrum segment [s dp -n d +1, s dp ], and the procedure is repeated for the rest of demands. As shown in Christodoulopoulos et al. 2011, MSF is an effective heuristic for minimizing Φ, which is our optimization objective. The application of MSF is straightforward in the SC scenario since the allocated spectrum segment is the same on both primary path p and backup path q, i.e., s dp =s dq. Therefore, in SC a criteria for selecting a pair of primary-backup paths minimization of s dp, i.e., from set Q d for demand d is the. In the DC scenario, the choice of a criteria function for selecting a path pair (p,q) from set Q d is more complicated since s dp and s dq may differ 7

8 M. Klinkowski and the optimization according to one of them does not necessarily lead to the minimization of the objective function (Φ). Therefore, we consider several alternative criteria functions for selecting a path pair in DC, namely: 1), 2), 3), 4), where path pair is selected as. In the Algorithm Tuning section, we show that, among these criteria, option 4) leads to best results. EA Details Figure 1. EA components: a) chromosome, b) crossover operator, c) mutation operator. In our EA, each individual in the population represents a sequence (permutation) of demands. The objective of EA is to find such a permutation of demands which, when processed by the RSA subroutine, minimizes the occupied spectrum width. The EA algorithm starts with creating the initial population. Better permutations are found over a number of generations. Operators such as crossover and mutation explore further possible permutations. In our approach, all the individuals are feasible solutions therefore no additional constraint handling procedures are necessary. When the stopping criterion is satisfied, a near-optimal sequence of demands is found. Below we describe in details the relevant building blocks of our EA. Representation: The encoded solution (chromosome) is a vector of numbers, coded as π=[π 1,,π D ], which represents a sequence of demands (to be processed by the RSA algorithm). In order to assure that the demands are sorted in decreasing order of n d, which is required by our MSF-based RSA algorithm, vector π of each chromosome is formed and maintained by EA as a concatenation of smaller vector segments π n, where π n represents a permutation vector of demands belonging to D n for 8

9 Evolutionary Algorithm for DPP in EON each n N. As a result, starting with segment π nmax, which is a permutation vector of demands belonging to D nmax, we form π=[π nmax,,π n,,π nmin ]. In Figure 1a) we present two exemplary chromosomes p1 and p2 formed for the case of D =9 demands among which there are 3 demands requesting 6 slices (segment π 6, in position 1-3), 2 demands requesting 4 slices (segment π 4, in position 4-5), and 4 demands requesting 2 slices (segment π 2, in position 6-9). The order of demands is ( ) and ( ), respectively, for p1 and p2. Population Initialization: The EA starts with an initial population of chromosomes. Permutation vectors π n, n N, of each chromosome are generated randomly. As discussed in Talbi 2009, the larger the population the better the convergence toward good results, however, at the cost of time complexity. Therefore, in this work, a population of M=60 chromosomes is maintained, which is a practical value as stated in Talbi Fitness Function: The objective function to be minimized is the spectrum width, which is an output of our greedy RSA algorithm (described in the RSA Subroutine section). Therefore, in order to drive the search toward better solutions, we use a fitness value (f) that corresponds to the objective value (Φ), in particular, f = -Φ. Selection of Chromosomes for the Next Generation: We use the stochastic universal sampling method (Talbi 2009) for selecting two parents to produce new individuals for the next generation. Crossover: We use the uniform crossover operator (Talbi 2009) by which a single offspring is created from two parents by coping randomly segments π n from either parent for each n N. In Figure 1b), segments π 6 and π 2 are selected from parent p1 and segment π 4 is selected from parent p2 to create offspring o. Mutation: The mutation procedure is applied after the crossover on each offspring independently. Specifically, for randomly selected segment π n, n N, we exchange the position of two randomly 9

10 M. Klinkowski selected elements in this segment. In Figure 1c), segment π 2 is selected and, within this segment, the position of demands 9 and 6 is exchanged. Replacement: New individuals replace old individuals if their fitness values are better than those of the old ones. Apart from that, in this work we consider that: a) at least E best individuals survive to the next generation, and b) X percent of the population at the next generation, not including the survivals from the old population, are created by the crossover function. In the Algorithm Tuning section we evaluate the impact of E and X on the algorithm performance. The population size at each generation is maintained fixed and equal to M. In the Numerical Result section we evaluate different values of M. Stopping Criteria: Our EA terminates after G generations. In this work, we perform the algorithm evaluation for G {1,10,50,100,200}. Algorithm Complexity The algorithm complexity is polynomial in time. In particular, the computation complexity of MSF is bounded by O( D k S E ), where D is the number of demands to be processed, k is the number of candidate path pairs per demand, and S E corresponds to the (worst-case) complexity of the search of unoccupied slices. MSF is called once for each chromosome. The number of chromosomes is limited by the product of the population size (M) and the number of generations (G) in EA. NUMERICAL RESULTS In this section, we evaluate the performance of our EA-MSF algorithm, presented in the previous section, and compare it with other reference algorithms. We focus on the occupied spectrum width (Φ), in terms of the number of slices, and computation time (T; in seconds). As the reference algorithms we use: AG - an Auxiliary Graph-based algorithm proposed in Patel et al for dedicated path protection, and a pure MSF algorithm. Note that MSF was originally developed in the context of the RSA problem without additional survivability constraints. Therefore, we have adapted it in order to address the SC and DC constraints which are present in the DPP scenario (see the RSA 10

11 Evolutionary Algorithm for DPP in EON Subroutine section for implementation details). Moreover, for small network scenarios, we provide optimal results obtained with an ILP formulation of the RSA/DPP problem (we refer to Klinkowski and Walkowiak 2012 for the formulation). Also, we provide a lower bound (LB) on Φ, obtained by solving a multicommodity routing problem with the spectrum continuity and channel assignment constraints relaxed (see Sec. III-A in Christodoulopoulos et al. 2011). Figure 2. Network topologies. The evaluation is performed for SIMPLE6 (6 nodes, 16 links), INT9 (9 nodes, 26 links), NSF15 (15 nodes, 46 links), and UBN24 (24 nodes, 86 links) network topologies (see Figure 2). We set S =1500, which is large enough to accommodate all demands without their blocking in the most demanding scenario; in other words, we assure that Φ S. Candidate pairs of primary-backup paths are link disjoint paths and they are calculated as the shortest paths, taking into account the overall length of both paths. We consider k {2,3, 10,30}. The requested spectrum n d is an even number generated with a uniform distribution and for randomly selected demand pair (o d, t d ), where o d,t d V and o d t d. In particular, n d N={2,4,..., n max }, where n max {8,16}. For fair evaluation, we consider all algorithms operate on an ordered set of demands D (as defined in the previous section). Both the EA procedure of EA-MSF and the AG algorithm are implemented in MATLAB. In order to speed up the processing of the MSF subroutine, we implement it in C++ and call as an external function from MATLAB. We use IBM ILOG CPLEX v.12.4 to solve ILP. The evaluation is performed 11

12 M. Klinkowski on an Intel i5 3.3GHz 16GB computer. If not mentioned otherwise, the results are averaged over 100 randomly generated demand sets. Algorithm Tuning We begin with comparing different criteria functions to be used for selecting routes in the MSF algorithm in the DC scenario (see the discussion in the RSA Subroutine section). In Table 1, we present the corresponding evaluation results obtained in the NSF15 and UBN24 networks under, respectively, D =210 and D =552, and n max =16 for both networks. As we can see, the minimization of the sum of slice indexes on both primary and backup paths (i.e., min{s dp +s dq }) leads to best results. Therefore, in the remainder of the paper we consider that this criteria is applied in the DC scenario in both the MSF algorithm and the RSA subroutine of EA-MSF. TABLE 1 Performance of MSF with different route selection criteria functions in DC. Scenario min{s dp } min{s dq } min{max{s dp,s dq }} min{s dp + s dq } k Φ NSF ,98 463,66 464,94 462, ,50 416,34 400,26 393, ,46 393,34 381,40 377,12 UBN , , ,24 999, ,24 812,28 794,72 739, ,46 738,38 731,02 705,76 Next, we evaluate the impact of two EA parameters, namely, E and X, on the performance of EA-MSF algorithm. The evaluation if performed in the NSF15 network with D =210, n max =16, k=10 and under the SC constraint. We consider E {2,4,6,12}, and X {20%,30%,50%,80%}. In Table 2, we present relative results of Φ with respect to the minimum value of Φ obtained for E=2 and X=50%. In the remainder of the paper, we evaluate EA-MSF with E=2 and X=50%. TABLE 2 Performance of EA-MSF under different values of E and X. E \ X 20% 30% 50% 80% Average 2 0,11% 0,17% 0,00% 0,70% 0,24% 4 0,16% 0,45% 0,16% 0,55% 0,33% 6 0,10% 0,30% 0,41% 0,78% 0,40% 12 0,41% 0,47% 0,17% 1,09% 0,53% 12

13 Evolutionary Algorithm for DPP in EON Average 0,19% 0,35% 0,19% 0,78% Optimality of EA-MSF In Table 3 we compare performance results of heuristic algorithms with the optimal ones (ILP) in small network scenarios. We consider D =10 and D =15, respectively, for SIMPLE6 and INT9, and n max =8 for both networks. EA-MSF runs G=50 generations. Due to the complexity of ILP, which does not reach optimality for about 62 out of 150 cases (39%) in a 2h period, the results for INT9 are averaged over the remaining 88 demand sets. Note that LB is the same for both SC and DC scenarios since in the multicommodity routing problem, which is solved to achieve the lower bound, the channel assignment constraints responsible for the difference between SC and DC are relaxed. We can see that EA-MSF outperforms other heuristics and its performance gap to optimal results is between 0,80-3,60%. Also, we can see that MSF and, consequently, EA-MSF perform better in the SC scenario than in the DC scenario. It comes from the fact that RSA decisions in the original version of MSF concern single route selections, while in DPP a pair of paths has to be selected. Now, since the selection of a disjoint pair of paths (p,q) with a minimum slice index in SC can be considered as a selection of super-path P=p q with the minimum slice index (see the RSA Subroutine section), therefore adapted MSF performs well in the SC scenario. However, the case of DC is not trivial. Here, both p and q may occupy different channels and, therefore, the adapted version of MSF requires route selection strategies that incorporate the slice indexes of both paths. As we show in the Algorithm Tuning section, a good strategy leading to minimal Φ results is to minimize the sum of the slice indexes. Eventually, the development of more sophisticated RSA algorithms for DC will be required so that to find solutions closer to optimal ones and such investigation is left for future work. TABLE 3 Optimality gap. Scenario LB ILP AG MSF EA-MSF k Channel Φ Φ T Optimality gap SIMPLE6 2 SC 25,02 27, ,78% 3,96% 0,81% 13

14 M. Klinkowski 2 DC 25,02 25, ,28% 3,56% 3 SC 24,88 27, ,85% 3,45% 0,91% 3 DC 24,88 25, ,92% 3,60% INT9 2 SC 33,61 34,79 22,6 10,89% 3,80% 1,34% 2 DC 33,61 33,61 9,4-4,08% 2,04% Performance Comparison In Table 4 we present performance results obtained with heuristic algorithms in larger networks. In the evaluation, D =210 and D =552, respectively, for NSF15 and UBN24, and n max =16 for both networks. EA-MSF runs G=200 generations. Since AG is designed for the SC scenario and it calculates routing paths online, i.e., without using candidate paths, we present its results in a dedicated row. In the comparison AG vs. EA-MSF we consider k=30 for EA-MSF. Similarly as in small networks, EA-MSF achieves better performance (in terms of Φ) than the reference algorithms. In details, the performance gap, defined as a relative performance improvement, is of about 20%-35% and 4%-11%, respectively, when comparing EA-MSF with AG and MSF. We can see that Φ can be improved considerably (about 15-19% for NSF15 and 20%-30% for UBN24) if there are more path pairs available in the set of candidate paths (k=2 vs. k=30). For MSF, the average computation times are below 0,5 seconds in the most demanding scenario. The computation times for EA-MSF are presented in the next section. TABLE 4 Comparison of heuristic algorithms. Scenario LB AG MSF EA-MSF EA-MSF vs. AG EA-MSF vs. MSF k Channel Φ Performance gap NSF15 2 SC 428,5-537,8 492,1-9,30% 2 DC 428,5-462,8 434,61-6,49% 10 SC 333,2-485,4 436,7-11,16% 10 DC 333,2-393,6 364,0-8,10% 30 SC 321,3-461,96 417,2-10,71% 30 DC 321,3-377,1 352,5-6,98% - SC - 566, ,79% - UBN24 2 SC 955,7-1083,3 1026,9-5,49% 2 DC 955,7-999,1 963,5-3,70% 10 SC 665,6-923,5 856,6-7,81% 10 DC 665,6-739,5 703,9-5,06% 14

15 Evolutionary Algorithm for DPP in EON 30 SC 665,3-836,6 785,5-6,50% 30 DC 665,3-705,8 681,0-3,56% - SC - 942, ,96% - EA-MSF Performance In Table 5 we present the EA-MSF performance results in a function of the number of generations (G). The improvement between G=1 and G=200 is of about 1%-5%, however, at the cost of extended computation time, which grows almost linearly with G. The impact of k on computation times is much smaller. Therefore, it is worth to have a large set of candidate paths since it has a great impact on the value of Φ (as shown in the previous section). We can see that above certain G the objective function (Φ) does not improve significantly. In particular, we find that G=50 is a reasonable trade-off between solution quality and computation time for EA-MSF in the studied scenarios. In order to evaluate the effectiveness of the evolutionary-base technique applied in EA-MSF, we have compared it with a random-based algorithm (RND). RND runs MSF for a number (denoted as K) of randomly generated permutations of the demand set and selects the solution with the best objective function. For the sake of comparability, we have assumed K=G*M iterations, which is roughly the total number of individuals generated during the EA-MSF algorithm runtime; in the evaluation, M=50 and, therefore, K=3000. In the analyzed large network scenarios with k {2,10,30}, in average, EA-MSF always provide better results than RND. We report that the performance gap of EA-MSF vs. RND is of about 1,90%-2,40% in NSF15 and 1,00%-1,50% in UBN24. TABLE 5 Performance of EA-MSF in a function of the number of generations (G); T in seconds. Scenario G = 1 G = 10 G = 50 G = 100 G = 200 G = 1 vs. G = 200 k Channel Φ T Φ T Φ T Φ T Φ T Improvement in Φ NSF15 2 SC 513 1, , ,33% 2 DC 443 1, , ,95% 10 SC 459 1, , ,16% 10 DC 376 2, ,36% 30 SC 436 2, ,61% 30 DC 363 3, ,98% UBN24 15

16 M. Klinkowski 2 SC , ,27% 2 SC 976 7, ,26% 10 SC ,60% 10 DC ,23% 30 SC ,14% 30 DC ,45% Comparison of DPP scenarios Eventually, we discuss a trade-off between having less spectrum resources occupied in the network at the cost of applying a more complex DC scenario. To support the discussion, in Table 6 we present comparative results of the spectrum usage in both SC and DC survivable network scenarios. The results are obtained with both MSF and EA-MSF algorithms. We can see that in all reported cases the SC scenario requires at least 15% more spectrum resources than the DC scenario. It is a consequence of the constraint to have both primary and backup lightpaths allocated within the same subset of slices in SC. Since higher spectrum requirements can be translated to higher network costs, as discussed in Palkopoulou et al. 2012, there is a rationale to consider the DC scenario as a valid scenario for EON. However, further analysis, taking into account the cost of re-configurable and tunable transponders, will be required in order to evaluate the cost impact of applying the DC scenario in EON. TABLE 6 Comparison of DPP scenarios; k=30. Network Algorithm Φ in SC Φ in DC Difference in Spectrum Usage NSF15 MSF EA-MSF ,5% 18,4% UBN24 MSF EA-MSF ,5% 15,3% CONCLUSIONS In this paper, we have focused on the problem of off-line routing and spectrum allocation in survivable elastic optical networks with dedicated path protection. To solve the problem, we have proposed a novel EA-MSF algorithm which is based on the evolutionary algorithm metaheuristic combined with a greedy RSA algorithm - as the RSA algorithm we use the MSF algorithm. The aim of EA is to help in the search of RSA solutions which minimize the occupied spectrum width. 16

17 Evolutionary Algorithm for DPP in EON To assess the algorithm performance, we have compared it with reference RSA/DPP algorithms. The performed numerical experiments show that EA-MSF outperforms other reference heuristics. Although the joint application of EA with the MSF algorithm leads to near-optimal results in the same channel allocation scenario, still, in the different channel allocation scenario in which the MSF algorithm has some difficulties, more sophisticated methods are required. Moreover, we have shown that the DC approach allows to save at least 13% of the overall spectrum width than when allocating identical channels to primary and backup lightpaths. This gain is at the expense of tunable transponders that have to be used in the DC scenario instead of shared and simple ones in the SC scenario. In future research, we plan to investigate the application of large-scale optimization methods, such as decomposition methods, for solving the RSA-DPP problem as well as to consider other survivability scenarios in EON including the use of shared protection. REFERENCES Assis, K. D. R., Almeida, R. C., and Waldman, H MILP formulation for squeezed protection in spectrum-sliced elastic optical path networks. Proceedings of the International Symposium on Performance Evaluation of Computer and Telecommunication Systems (SPECTS 2012). Blum, R Evolving Tunable Technologies in Transponder and Transceiver Modules. Technologies-small.pdf. Buus, J. and Murphy, E. J Tunable Lasers in Optical Networks. IEEE Journal of Lightwave Technology 24(1): Castro, A., Ruiz, M., Velasco, L., Junyent, G., and Comellas, J. Path-based recovery in flexgrid optical networks. Proceedings of International Conference on Transparent Optical Networks (ICTON 2012). Christodoulopoulos, K., Tomkos, I., and Varvarigos, E Elastic bandwidth allocation in flexible OFDM based optical networks. IEEE Journal of Lightwave Technology 29(9):

18 M. Klinkowski Palkopoulou, E., Angelou, M., Klonidis, D., Christodoulopoulos, K., Klekamp, A., Buchali, F., Varvarigos, E., and Tomkos, I Quantifying spectrum, cost, and energy efficiency in fixedgrid and flex-grid networks. IEEE/OSA Journal of Optical Communications and Networking 4(11): B42-B51. Cugini, F., Meloni, G., Paolucci, F., Sambo, N., Secondini, M., Gerardi, L., Poti, L., and Castoldi, P Demonstration of flexible optical network based on path computation element. IEEE Journal of Lightwave Technology 30(5): Eira, A., Pedro, J., and Pires, J Optimized design of shared restoration in flexible-grid transparent optical networks. Proceedings of the Optical Fiber Communication Conference (OFC 2012). Geisler, D.J., Proietti R., Yin, Y., Scott, R. P., Cai, X., Fontaine, N. K., Paraschis, L., Gerstel, O., and Yoo, S. J. B The first testbed bemonstration of a flexible bandwidth network with a realtime adaptive control plane. Proceedings of the European Conference on Optical Communication (ECOC 2011). Gerstel, O., Jinno M., and Lord, A Elastic optical networking: A new dawn for the optical layer? IEEE Communications Magazine 50(2): Gong, L., Zhou, X., Lu, W., and Zhu, Z A Two-Population Based Evolutionary Approach for Optimizing Routing, Modulation and Spectrum Assignments (RMSA) in O-OFDM Networks. IEEE Communications Letters 16(9): IBM IBM ILOG CPLEX optimizer. ITU-T G (ed. 2.0) Recommendation. Spectral grids for WDM applications: DWDM frequency grid. February Jinno, M., Takara H., Kozicki B., Tsukishima Y., Sone Y., and Matsuoka S Spectrum-efficient and scalable elastic optical path network: Architecture, benefits, and enabling technologies. IEEE Communications Magazine 47(11):

19 Evolutionary Algorithm for DPP in EON Klinkowski, M. and Walkowiak, K Routing and spectrum assignment in spectrum sliced elastic optical path network. IEEE Communications Letters 15(8): Klinkowski, M., Walkowiak, K., and Jaworski, M Off-line algorithms for routing, modulation level, and spectrum assignment in elastic optical networks. Proceedings of the IEEE International Conference on Transparent Optical Networks (ICTON 2011). Klinkowski, M A genetic algorithm for solving RSA problem in elastic optical networks with dedicated path protection. Proceedings of the International Conference on Computational Intelligence in Security for Information Systems (CISIS 2012). Klinkowski, M. and Walkowiak, K Offline RSA algorithms for elastic optical networks with dedicated path protection consideration. Proceedings of the International Workshop on Reliable Networks Design and Modeling (RNDM 2012). Liu, M., Tornatore, M., and Mukherjee, B Survivable Traffic Grooming in Elastic Optical Networks - Shared Path Protection. Proceedings of the International Conference on Communications (ICC 2012). Meloni, G., Paolucci, F., Sambo, N., Cugini, F., Secondini, M., Gerardi, L., Poti, L., and Castoldi, P PCE architecture for flexible WSON enabling dynamic rerouting with modulation format adaptation. Proceedings of the European Conference on Optical Communication (ECOC 2011). Patel, A.N., Ji, P. N., Jue, J. P., Wang, T Survivable transparent flexible optical WDM (FWDM) networks. Proceedings of the Optical Fiber Communication Conference (OFC 2011). Shao, X., Yeo, Y.-K., Xu, Z., Cheng, X., and Zhou, L., 2012: Shared-path protection in OFDM-based optical networks with elastic bandwidth allocation. Proceedings of the Optical Fiber Communication Conference (OFC 2012). Takagi, T., Hasegawa, H., Sato, K., Tanaka, T., Kozicki, B., Sone, Y., and Jinno, M Algorithms for maximizing spectrum efficiency in elastic optical path networks that adopt distance adaptive modulation. Proceedings of the European Conference on Optical Communication (ECOC 2010). 19

20 M. Klinkowski Talbi, E.G Metaheuristics: From Design to Implementation. Wiley. Velasco, L., Klinkowski, M., Ruiz, M., Comellas, J Modeling the routing and spectrum allocation problem for flexgrid optical networks. Photonic Network Communications 24(3): Wei, W., Wang, Ch., and Yu, J Cognitive optical networks: Key drivers, enabling techniques, and adaptive bandwidth services. IEEE Communications Magazine 50(1):

Optimization Algorithms for Data Center Location Problem in Elastic Optical Networks

Optimization Algorithms for Data Center Location Problem in Elastic Optical Networks Optimization Algorithms for Data Center Location Problem in Elastic Optical Networks Mirosław Klinkowski 1, Krzysztof Walkowiak 2, and Róża Goścień 2 1 National Institute of Telecommunications, Warsaw,

More information

Survivable Multipath Provisioning in OFDM-Based Flexible Optical Networks

Survivable Multipath Provisioning in OFDM-Based Flexible Optical Networks Survivable Multipath Provisioning in OFDM-Based Flexible Optical Networks Nan Xiao and Lu Ruan Department of Computer Science Iowa Stat University, Ames, IA 11 Email: {nxiao, ruanlu}@iastate.edu Abstract

More information

Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic

Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic A. Asensio 1, M. Klinkowski 2, M. Ruiz 1, V. López 3, A. Castro 1, L. Velasco 1, J. Comellas 1

More information

Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic

Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic Impact of Aggregation Level on the Performance of Dynamic Lightpath Adaptation under Time-Varying Traffic A. Asensio 1, M. Klinkowski 2, M. Ruiz 1, V. López 3, A. Castro 1, L. Velasco 1, J. Comellas 1

More information

Spectrum Allocation Policies in Fragmentation Aware and Balanced Load Routing for Elastic Optical Networks

Spectrum Allocation Policies in Fragmentation Aware and Balanced Load Routing for Elastic Optical Networks Spectrum Allocation Policies in Fragmentation Aware and Balanced Load Routing for Elastic Optical Networks André C. S. Donza, Carlos R. L. Francês High Performance Networks Processing Lab - LPRAD Universidade

More information

Dynamic Routing and Resource Allocation in a Elastic Optical Network Using Learning Algorithms. Tanjila Ahmed

Dynamic Routing and Resource Allocation in a Elastic Optical Network Using Learning Algorithms. Tanjila Ahmed Dynamic Routing and Resource Allocation in a Elastic Optical Network Using Learning Algorithms Tanjila Ahmed Contents Motivation Novelties of Elastic Optical Network Enabling Technology for Elastic Optical

More information

On Efficient Protection Design for Dynamic Multipath Provisioning in Elastic Optical Networks

On Efficient Protection Design for Dynamic Multipath Provisioning in Elastic Optical Networks On Efficient Protection Design for Dynamic Multipath Provisioning in Elastic Optical Networks Xiaoliang Chen, Shilin Zhu,DiChen, Shuguang Hu, Chang Li, Zuqing Zhu School of Information Science and Technology,

More information

Dynamic Routing and Spectrum Assignment Algorithms in Flexible Optical Networks: An Overview

Dynamic Routing and Spectrum Assignment Algorithms in Flexible Optical Networks: An Overview 2012 7th International ICST Conference on Communications and Networking in China (CHINACOM) Dynamic Routing and Spectrum Assignment Algorithms in Flexible Optical Networks: An Overview Nan Hua*, Yang Liu,

More information

Optimization of Spectrally and Spatially Flexible Optical Networks with Spatial Mode Conversion

Optimization of Spectrally and Spatially Flexible Optical Networks with Spatial Mode Conversion 148 Regular papers ONDM 2018 Optimization of Spectrally and Spatially Flexible Optical Networks with Spatial Mode Conversion Mirosław Klinkowski, Grzegorz Zalewski, Krzysztof Walkowiak, National Institute

More information

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 33, NO. 1, JANUARY 1,

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 33, NO. 1, JANUARY 1, JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 33, NO. 1, JANUARY 1, 2015 275 Dynamic Multipath Routing With Traffic Grooming in OFDM-Based Elastic Optical Path Networks Zheyu Fan, Student Member, IEEE, Yang Qiu,

More information

Performance Evaluation of Non-Hitless Spectrum Defragmentation Algorithms in Elastic Optical Networks

Performance Evaluation of Non-Hitless Spectrum Defragmentation Algorithms in Elastic Optical Networks Performance Evaluation of Non-Hitless Spectrum Defragmentation Algorithms in Elastic Optical Networks Sergio Fernández Martínez and Diego P. Pinto-Roa Facultad Politécnica Universidad Nacional de Asunción

More information

An Efficient Algorithm for Solving Traffic Grooming Problems in Optical Networks

An Efficient Algorithm for Solving Traffic Grooming Problems in Optical Networks An Efficient Algorithm for Solving Traffic Grooming Problems in Optical Networks Hui Wang, George N. Rouskas Operations Research and Department of Computer Science, North Carolina State University, Raleigh,

More information

Migration Towards Terascale Flexible-Grid Optical Networks. Tanjila Ahmed

Migration Towards Terascale Flexible-Grid Optical Networks. Tanjila Ahmed Migration Towards Terascale Flexible-Grid Optical Networks Tanjila Ahmed Motivation With traffic growing at 40 percent annually, optical spectrum is becoming scarce, network architectures supporting channels

More information

Bandwidth Defragmentation in Dynamic Elastic Optical Networks with Minimum Traffic Disruptions

Bandwidth Defragmentation in Dynamic Elastic Optical Networks with Minimum Traffic Disruptions IEEE ICC 2013 - Optical Networks and Systems Bandwidth Defragmentation in Dynamic Elastic Optical Networks with Minimum Traffic Disruptions Mingyang Zhang, Weiran Shi, Long Gong, Wei Lu, Zuqing Zhu School

More information

Dynamic Service Provisioning of Advance Reservation Requests in Elastic Optical Networks

Dynamic Service Provisioning of Advance Reservation Requests in Elastic Optical Networks JOURNAL OF LIGHTWAVE TECHNOLOGY 1621 Dynamic Service Provisioning of Advance Reservation Requests in Elastic Optical Networks Wei Lu, Student Member, IEEE, and Zuqing Zhu, Senior Member, IEEE Abstract

More information

PATH SPLITTING FOR VIRTUAL NETWORK EMBEDDING IN ELASTIC OPTICAL NETWORKS

PATH SPLITTING FOR VIRTUAL NETWORK EMBEDDING IN ELASTIC OPTICAL NETWORKS PATH SPLITTING FOR VIRTUAL NETWORK EMBEDDING IN ELASTIC OPTICAL NETWORKS Badr Oulad Nassar and Takuji Tachibana Graduate School of Engineering, University of Fukui, Fukui City, Japan ABSTRACT In elastic

More information

Optimal Design for Shared Backup Path Protected Elastic Optical Networks Under Single-Link Failure

Optimal Design for Shared Backup Path Protected Elastic Optical Networks Under Single-Link Failure Shen et al. VOL. 6, NO. 7/JULY 2014/J. OPT. COMMUN. NETW. 649 Optimal Design for Shared Backup Path Protected Elastic Optical Networks Under Single-Link Failure Gangxiang Shen, Yue Wei, and Sanjay K. Bose

More information

Toward the joint design of electronic and optical layer protection

Toward the joint design of electronic and optical layer protection Toward the joint design of electronic and optical layer protection Massachusetts Institute of Technology Slide 1 Slide 2 CHALLENGES: - SEAMLESS CONNECTIVITY - MULTI-MEDIA (FIBER,SATCOM,WIRELESS) - HETEROGENEOUS

More information

A Novel Class-based Protection Algorithm Providing Fast Service Recovery in IP/WDM Networks

A Novel Class-based Protection Algorithm Providing Fast Service Recovery in IP/WDM Networks A Novel Class-based Protection Algorithm Providing Fast Service Recovery in IP/WDM Networks Wojciech Molisz and Jacek Rak Gdansk University of Technology, G. Narutowicza 11/12, Pl-8-952 Gdansk, Poland

More information

Multi-Area Centralized Control Plane of an EON

Multi-Area Centralized Control Plane of an EON Multi-Area Centralized Control Plane of an EON Eric Sturzinger Group Meeting JUL 0 Overview Objectives / Problem Statement Previous Work Decentralized Control Plane (GMPLS) Centralized Control Plane (SDN)

More information

How to overcome the capacity crunch new challenges in optical networks

How to overcome the capacity crunch new challenges in optical networks How to overcome the capacity crunch new challenges in optical networks Krzysztof Walkowiak Department of Systems and Computer Networks Wroclaw University of Technology, Wrocław, Poland February 4, 2016,

More information

Adaptation and Monitoring for Elastic Alien Wavelengths [Invited]

Adaptation and Monitoring for Elastic Alien Wavelengths [Invited] Adaptation and Monitoring for Elastic Alien Wavelengths [Invited] F. Cugini CNIT Pisa, Italy Email: filippo.cugini@cnit.it N. Sambo, F. Paolucci, F. Fresi, P. Castoldi Scuola Superiore Sant Anna Pisa,

More information

A Novel Genetic Approach to Provide Differentiated Levels of Service Resilience in IP-MPLS/WDM Networks

A Novel Genetic Approach to Provide Differentiated Levels of Service Resilience in IP-MPLS/WDM Networks A Novel Genetic Approach to Provide Differentiated Levels of Service Resilience in IP-MPLS/WDM Networks Wojciech Molisz, DSc, PhD Jacek Rak, PhD Gdansk University of Technology Department of Computer Communications

More information

Migration Steps Towards Flexi-grid Networks

Migration Steps Towards Flexi-grid Networks Future Network & MobileSummit 2013 Conference Proceedings Paul Cunningham and Miriam Cunningham (Eds) IIMC International Information Management Corporation, 2013 ISBN: 978-1-905824-36-6 Migration Steps

More information

Dynamic RMSA in Spectrum-Sliced Elastic Optical Networks for High-Throughput Service Provisioning

Dynamic RMSA in Spectrum-Sliced Elastic Optical Networks for High-Throughput Service Provisioning 2013 International Conference on Computing, Networking and Communications, Optical and Grid Networking Symposium Dynamic RMSA in Spectrum-Sliced Elastic Optical Networks for High-Throughput Service Provisioning

More information

A Novel Optimization Method of Optical Network Planning. Wu CHEN 1, a

A Novel Optimization Method of Optical Network Planning. Wu CHEN 1, a A Novel Optimization Method of Optical Network Planning Wu CHEN 1, a 1 The engineering & technical college of chengdu university of technology, leshan, 614000,china; a wchen_leshan@126.com Keywords:wavelength

More information

OFC/NFOEC 12 Summary --Elastic Networks. Avishek Nag

OFC/NFOEC 12 Summary --Elastic Networks. Avishek Nag OFC/NFOEC 12 Summary --Elastic Avishek Nag OFC 12 Technical Sessions with Papers Related to Elastic OW3A: Impairment-Aware Networking : Elastic Optical NW3F: ROADM Network Architectures OTh4B: Next Generation

More information

Optical Communications and Networking 朱祖勍. Nov. 27, 2017

Optical Communications and Networking 朱祖勍. Nov. 27, 2017 Optical Communications and Networking Nov. 27, 2017 1 What is a Core Network? A core network is the central part of a telecommunication network that provides services to customers who are connected by

More information

Network Topology Control and Routing under Interface Constraints by Link Evaluation

Network Topology Control and Routing under Interface Constraints by Link Evaluation Network Topology Control and Routing under Interface Constraints by Link Evaluation Mehdi Kalantari Phone: 301 405 8841, Email: mehkalan@eng.umd.edu Abhishek Kashyap Phone: 301 405 8843, Email: kashyap@eng.umd.edu

More information

Design of Large-Scale Optical Networks Λ

Design of Large-Scale Optical Networks Λ Design of Large-Scale Optical Networks Λ Yufeng Xin, George N. Rouskas, Harry G. Perros Department of Computer Science, North Carolina State University, Raleigh NC 27695 E-mail: fyxin,rouskas,hpg@eos.ncsu.edu

More information

A Scalable Wavelength Assignment Algorithm Using Minimal Number of Wavelength Converters in Resilient WDM Networks

A Scalable Wavelength Assignment Algorithm Using Minimal Number of Wavelength Converters in Resilient WDM Networks A Scalable Wavelength Assignment Algorithm Using Minimal Number of Wavelength Converters in Resilient WDM Networks Technical Report UTD/EE/4/2009 September 2009 Miguel Razo, Shreejith Billenahalli, Wanjun

More information

Design Methodologies and Algorithms for Survivable C-RAN

Design Methodologies and Algorithms for Survivable C-RAN 16 Regular papers ONDM 218 Design Methodologies and Algorithms for Survivable C-RAN Bahare M. Khorsandi, Federico Tonini, Carla Raffaelli DEI, University of Bologna Viale Risorgimento 2, 4136 Bologna,

More information

Bit Index Explicit Replication (BIER) Multicasting in Transport Networks [Invited]

Bit Index Explicit Replication (BIER) Multicasting in Transport Networks [Invited] Bit Index Explicit Replication (BIER) Multicasting in Transport Networks [Invited] A. Giorgetti, A. Sgambelluri, F. Paolucci, N. Sambo, P. Castoldi Scuola Superiore Sant Anna Pisa, Italy Email: a.giorgetti@santannapisa.it

More information

Hierarchical Traffic Grooming Formulations

Hierarchical Traffic Grooming Formulations Hierarchical Traffic Grooming Formulations Hui Wang, George N. Rouskas Operations Research and Department of Computer Science, North Carolina State University, Raleigh, NC 27695-8206 USA Abstract Hierarchical

More information

Analysis and Algorithms for Partial Protection in Mesh Networks

Analysis and Algorithms for Partial Protection in Mesh Networks Analysis and Algorithms for Partial Protection in Mesh Networks Greg uperman MIT LIDS Cambridge, MA 02139 gregk@mit.edu Eytan Modiano MIT LIDS Cambridge, MA 02139 modiano@mit.edu Aradhana Narula-Tam MIT

More information

An Integer Programming Approach to Packing Lightpaths on WDM Networks 파장분할다중화망의광경로패킹에대한정수계획해법. 1. Introduction

An Integer Programming Approach to Packing Lightpaths on WDM Networks 파장분할다중화망의광경로패킹에대한정수계획해법. 1. Introduction Journal of the Korean Institute of Industrial Engineers Vol. 32, No. 3, pp. 219-225, September 2006. An Integer Programming Approach to Packing Lightpaths on WDM Networks Kyungsik Lee 1 Taehan Lee 2 Sungsoo

More information

GMPLS control plane extensions in support of flex-grid enabled elastic optical networks

GMPLS control plane extensions in support of flex-grid enabled elastic optical networks Downloaded from orbit.dtu.dk on: Feb 08, 2018 GMPLS control plane extensions in support of flex-grid enabled elastic optical networks Turus, Ioan; Fagertun, Anna Manolova; Dittmann, Lars Published in:

More information

Who shot optical packet switching?

Who shot optical packet switching? Who shot optical packet switching? José Roberto de Almeida Amazonas, Germán Santos-Boada, and Josep Solé-Pareta Department of Computer Architecture, Technical University of Catalonia, Barcelona, Spain

More information

Cost Evaluation for Flexible-Grid Optical Networks

Cost Evaluation for Flexible-Grid Optical Networks Cost Evaluation for Flexible-Grid Optical Networks Jorge López Vizcaíno 1, 3, Yabin Ye 1, Víctor López 2, Felipe Jiménez 2, Raúl Duque 2, Peter M. Krummrich 3 1 Huawei Technologies Duesseldorf GmbH, European

More information

IO2654 Optical Networking. WDM network design. Lena Wosinska KTH/ICT. The aim of the next two lectures. To introduce some new definitions

IO2654 Optical Networking. WDM network design. Lena Wosinska KTH/ICT. The aim of the next two lectures. To introduce some new definitions IO2654 Optical Networking WDM network design Lena Wosinska KTH/ICT 1 The aim of the next two lectures To introduce some new definitions To make you aware about the trade-offs for WDM network design To

More information

Improving Link Spectrum Utilization in FlexGrid Optical Networks

Improving Link Spectrum Utilization in FlexGrid Optical Networks 2015 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising

More information

Survivability with P-Cycle in WDM Networks

Survivability with P-Cycle in WDM Networks Survivability with P-Cycle in WDM Networks K. Aparna 1, P. Ramya Krishna 2 JNTUA College of Engineering, Pulivendula Abstract: In this paper we discuss about the Pre-Configured survivability schemes. Network

More information

Spectrum Allocation Policies for Flex Grid Network with Data Rate Limited Transmission

Spectrum Allocation Policies for Flex Grid Network with Data Rate Limited Transmission Spectrum Allocation Policies for Flex Grid Network with Data Rate Limited Transmission Kruthika Lohith 1, Triveni C L 2, Dr. P.C Srikanth 3 1Malnad College of Engineering, Hassan, Karnataka 2 Asst Professor,

More information

An Ant Colony Optimization Implementation for Dynamic Routing and Wavelength Assignment in Optical Networks

An Ant Colony Optimization Implementation for Dynamic Routing and Wavelength Assignment in Optical Networks An Ant Colony Optimization Implementation for Dynamic Routing and Wavelength Assignment in Optical Networks Timothy Hahn, Shen Wan March 5, 2008 Montana State University Computer Science Department Bozeman,

More information

GA-Based Hybrid Algorithm for MBR Problem of FIPP p-cycles for Node Failure on Survivable WDM Networks

GA-Based Hybrid Algorithm for MBR Problem of FIPP p-cycles for Node Failure on Survivable WDM Networks GA-Based Hybrid Algorithm for MBR Problem of FIPP p-cycles for Node Failure on Survivable WDM Networks Der-Rong Din Department of Computer Science and Information Engineering, National Changhua University

More information

Delayed reservation decision in optical burst switching networks with optical buffers

Delayed reservation decision in optical burst switching networks with optical buffers Delayed reservation decision in optical burst switching networks with optical buffers G.M. Li *, Victor O.K. Li + *School of Information Engineering SHANDONG University at WEIHAI, China + Department of

More information

Offline Distance-Adaptive Routing and Spectrum Assignment (DA-RSA) in Rings

Offline Distance-Adaptive Routing and Spectrum Assignment (DA-RSA) in Rings Offline Distance-Adaptive Routing and Spectrum Assignment (DA-RSA) in Rings Sahar Talebi, Iyad Katib, George N. Rouskas Operations Research and Department of Computer Science, North Carolina State University,

More information

Link Selection Algorithms for Link-Based ILPs and Applications to RWA in Mesh Networks

Link Selection Algorithms for Link-Based ILPs and Applications to RWA in Mesh Networks Link Selection Algorithms for Link-Based ILPs and Applications to RWA in Mesh Networks Zeyu Liu, George N. Rouskas Department of Computer Science, North Carolina State University, Raleigh, NC 27695-8206,

More information

ELASTIC optical networking has been the subject of considerable

ELASTIC optical networking has been the subject of considerable JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 33, NO. 1, JANUARY 1, 2015 151 On Routing and Spectrum Assignment in Rings Sahar Talebi, Evripidis Bampis, Giorgio Lucarelli, Iyad Katib, and George N. Rouskas Abstract

More information

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 31, NO. 1, JANUARY 1,

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 31, NO. 1, JANUARY 1, JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 31, NO. 1, JANUARY 1, 2013 15 Dynamic Service Provisioning in Elastic Optical Networks With Hybrid Single-/Multi-Path Routing Zuqing Zhu, Senior Member, IEEE, Wei

More information

The Spectrum Assigment (SA) Problem in Optical Networks: A Multiprocesor Scheduling Perspective

The Spectrum Assigment (SA) Problem in Optical Networks: A Multiprocesor Scheduling Perspective The Spectrum Assigment (SA) Problem in Optical Networks: A Multiprocesor Scheduling Perspective Sahar Talebi, Evripidis Bampis, Giorgio Lucarelli,IyadKatib,GeorgeN.Rouskas Operations Research and Department

More information

Dynamic Unicast/Multicast-Capable RMSA in Elastic Optical Networks

Dynamic Unicast/Multicast-Capable RMSA in Elastic Optical Networks Dynamic Unicast/Multicast-Capable RMSA in Elastic Optical Networks Zuqing Zhu, Xiang Zhou, Long Gong, Wenwen Zhao, Wei Lu School of Information Science and Technology University of Science and Technology

More information

3350 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 29, NO. 21, NOVEMBER 1, Reach Adapting Algorithms for Mixed Line Rate WDM Transport Networks

3350 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 29, NO. 21, NOVEMBER 1, Reach Adapting Algorithms for Mixed Line Rate WDM Transport Networks 3350 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 29, NO. 21, NOVEMBER 1, 2011 Reach Adapting Algorithms for Mixed Line Rate WDM Transport Networks Konstantinos Christodoulopoulos, Konstantinos Manousakis, and

More information

Survivable Routing Problem in EONs with FIPP p-cycles Protection

Survivable Routing Problem in EONs with FIPP p-cycles Protection JOURNAL OF INFORMATION SCIENCE AND ENGINEERING XX, XXXX-XXXX (2016) Survivable Routing Problem in EONs with FIPP p-cycles Protection DER-RONG DIN Department of Computer Science and Information Engineering

More information

Dynamic connection establishment and network re-optimization in flexible optical networks

Dynamic connection establishment and network re-optimization in flexible optical networks Photon Netw Commun (2015) 29:307 321 DOI 10.1007/s11107-015-0500-8 Dynamic connection establishment and network re-optimization in flexible optical networks P. Soumplis 1,2 K. Christodoulopoulos 1,2 E.

More information

A Network Optimization Model for Multi-Layer IP/MPLS over OTN/DWDM Networks

A Network Optimization Model for Multi-Layer IP/MPLS over OTN/DWDM Networks A Network Optimization Model for Multi-Layer IP/MPLS over OTN/DWDM Networks Iyad Katib and Deep Medhi Computer Science & Electrical Engineering Department University of Missouri-Kansas City, USA {IyadKatib,

More information

Copyrights IEEE 2016 This is the accepted version of the invited paper entitled A Survey of Multi-layer Network Optimization

Copyrights IEEE 2016 This is the accepted version of the invited paper entitled A Survey of Multi-layer Network Optimization Accepted version of the manuscript submitted to the 20 th International Conference on Optical Network Design and Modelling (ONDM), May 2016. Copyrights IEEE 2016 This is the accepted version of the invited

More information

FLEXING NEXT GENERATION OPTICAL MUSCLES

FLEXING NEXT GENERATION OPTICAL MUSCLES FLEXING NEXT GENERATION OPTICAL MUSCLES A Perspective on Flexi-rate Innovation and True 400G From high-capacity data center connectivity to LTE-enabled mobility, the foundation of our modern communications

More information

WDM Network Provisioning

WDM Network Provisioning IO2654 Optical Networking WDM Network Provisioning Paolo Monti Optical Networks Lab (ONLab), Communication Systems Department (COS) http://web.it.kth.se/~pmonti/ Some of the material is taken from the

More information

Arc Perturbation Algorithms for Optical Network Design

Arc Perturbation Algorithms for Optical Network Design Applied Mathematical Sciences, Vol. 1, 2007, no. 7, 301-310 Arc Perturbation Algorithms for Optical Network Design Zbigniew R. Bogdanowicz Armament Research, Development and Engineering Center Building

More information

Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks

Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks Piotr Lechowicz, Krzysztof Walkowiak Wroclaw University of Science and Technology Mirosław Klinkowski ENGINE Centre and National

More information

Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks

Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks Selection of Spectral-Spatial Channels in SDM Flexgrid Optical Networks Piotr Lechowicz, Krzysztof Walkowiak, Mirosław Klinkowski Department of Systems and Computer Networks, Wrocław University of Science

More information

OPTICAL NETWORKS. Virtual Topology Design. A. Gençata İTÜ, Dept. Computer Engineering 2005

OPTICAL NETWORKS. Virtual Topology Design. A. Gençata İTÜ, Dept. Computer Engineering 2005 OPTICAL NETWORKS Virtual Topology Design A. Gençata İTÜ, Dept. Computer Engineering 2005 Virtual Topology A lightpath provides single-hop communication between any two nodes, which could be far apart in

More information

SCATTER SEARCH BASED METAHEURISTIC FOR ROBUST OPTIMIZATION OF THE DEPLOYING OF DWDM TECHNOLOGY ON OPTICAL NETWORKS WITH SURVIVABILITY

SCATTER SEARCH BASED METAHEURISTIC FOR ROBUST OPTIMIZATION OF THE DEPLOYING OF DWDM TECHNOLOGY ON OPTICAL NETWORKS WITH SURVIVABILITY Yugoslav Journal of Operations Research 15 (2005), Number 1, 65-77 SCATTER SEARCH BASED METAHEURISTIC FOR ROBUST OPTIMIZATION OF THE DEPLOYING OF DWDM TECHNOLOGY ON OPTICAL NETWORKS WITH SURVIVABILITY

More information

Available online at ScienceDirect

Available online at   ScienceDirect Available online at www.sciencedirect.com ScienceDirect Procedia Technology 0 ( 0 ) 900 909 International Conference on Computational Intelligence: Modeling, Techniques and Applications (CIMTA-0) Multicast

More information

Contribution to the Defragmentation of the Frequency Spectrum of Elastic Optical Networks by Using Independent Sets in a Graph

Contribution to the Defragmentation of the Frequency Spectrum of Elastic Optical Networks by Using Independent Sets in a Graph International Journal of Computer Science and Telecommunications [Volume 6, Issue 3, March 2015] 1 ISSN 2047-3338 Contribution to the Defragmentation of the Frequency Spectrum of Elastic Optical Networks

More information

Internet Traffic Characteristics. How to take care of the Bursty IP traffic in Optical Networks

Internet Traffic Characteristics. How to take care of the Bursty IP traffic in Optical Networks Internet Traffic Characteristics Bursty Internet Traffic Statistical aggregation of the bursty data leads to the efficiency of the Internet. Large Variation in Source Bandwidth 10BaseT (10Mb/s), 100BaseT(100Mb/s),

More information

RWA on Scheduled Lightpath Demands in WDM Optical Transport Networks with Time Disjoint Paths

RWA on Scheduled Lightpath Demands in WDM Optical Transport Networks with Time Disjoint Paths RWA on Scheduled Lightpath Demands in WDM Optical Transport Networks with Time Disjoint Paths Hyun Gi Ahn, Tae-Jin Lee, Min Young Chung, and Hyunseung Choo Lambda Networking Center School of Information

More information

A Heuristic Algorithm for Designing OTN Over Flexible-Grid DWDM Networks

A Heuristic Algorithm for Designing OTN Over Flexible-Grid DWDM Networks A Heuristic Algorithm for Designing OTN Over Flexible-Grid DWDM Networks João R. Santos1,3, António Eira1,2, and João Pires2,3 1 Coriant Portugal, Lisboa, Portugal. Instituto de Telecomunicações, Instituto

More information

Implementation of Genetic Algorithm for Combined Routing and Dimensioning for Dynamic WDM Networks

Implementation of Genetic Algorithm for Combined Routing and Dimensioning for Dynamic WDM Networks Implementation of Genetic Algorithm for Combined Routing and Dimensioning for Dynamic WDM Networks Bhuthesh H K 1, Triveni C L 2 1M.Tech student, Dept. Of ECE, MCE Hassan, Karnataka, India 2Assistant Professor,

More information

Network Protection Design for MPLS Networks

Network Protection Design for MPLS Networks Network Protection Design for MPLS Networks Gaurav Agrawal, Dijiang Huang, Deep Medhi Computer Science and Electrical Engineering Department University of Missouri-Kansas City, MO 64110, USA Computer Science

More information

A Novel Generic Graph Model for Traffic Grooming in Heterogeneous WDM Mesh Networks

A Novel Generic Graph Model for Traffic Grooming in Heterogeneous WDM Mesh Networks IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 11, NO. 2, APRIL 2003 285 A Novel Generic Graph Model for Traffic Grooming in Heterogeneous WDM Mesh Networks Hongyue Zhu, Student Member, IEEE, Hui Zang, Member,

More information

Energy Efficiency Analysis for Dynamic Routing in Optical Transport Networks

Energy Efficiency Analysis for Dynamic Routing in Optical Transport Networks Downloaded from orbit.dtu.dk on: Mar 02, 2018 Energy Efficiency Analysis for Dynamic Routing in Optical Transport Networks Vizcaíno, Jorge López; Ye, Yabin; Tafur Monroy, Idelfonso Published in: 2012 IEEE

More information

AN EVOLUTIONARY APPROACH TO DISTANCE VECTOR ROUTING

AN EVOLUTIONARY APPROACH TO DISTANCE VECTOR ROUTING International Journal of Latest Research in Science and Technology Volume 3, Issue 3: Page No. 201-205, May-June 2014 http://www.mnkjournals.com/ijlrst.htm ISSN (Online):2278-5299 AN EVOLUTIONARY APPROACH

More information

A Heuristic Algorithm for Designing Logical Topologies in Packet Networks with Wavelength Routing

A Heuristic Algorithm for Designing Logical Topologies in Packet Networks with Wavelength Routing A Heuristic Algorithm for Designing Logical Topologies in Packet Networks with Wavelength Routing Mare Lole and Branko Mikac Department of Telecommunications Faculty of Electrical Engineering and Computing,

More information

An optically transparent ultra high speed LAN-ring employing OTDM

An optically transparent ultra high speed LAN-ring employing OTDM An optically transparent ultra high speed LAN-ring employing OTDM K. Bengi, G. Remsak, H.R. van As Vienna University of Technology, Institute of Communication Networks Gusshausstrasse 25/388, A-1040 Vienna,

More information

WDM Network Provisioning

WDM Network Provisioning IO2654 Optical Networking WDM Network Provisioning Paolo Monti Optical Networks Lab (ONLab), Communication Systems Department (COS) http://web.it.kth.se/~pmonti/ Some of the material is taken from the

More information

Index Terms Backbone Networks Optimization, Resilience, Virtual Private Network, Traffic Demands and Traffic Flows

Index Terms Backbone Networks Optimization, Resilience, Virtual Private Network, Traffic Demands and Traffic Flows International Journal of Computer Science and Telecommunications [Volume 2, Issue 8, November 2011] 12 ISSN 2047-3338 Designing Different Models of Robust IP/MPLS over DWDM Networks Marko Rožman University

More information

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 21, NO. 7, SEPTEMBER

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 21, NO. 7, SEPTEMBER IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 21, NO. 7, SEPTEMBER 2003 1173 A Comprehensive Study on Next-Generation Optical Grooming Switches Keyao Zhu, Student Member, IEEE, Hui Zang, Member,

More information

An Evolutionary Algorithm for the Multi-objective Shortest Path Problem

An Evolutionary Algorithm for the Multi-objective Shortest Path Problem An Evolutionary Algorithm for the Multi-objective Shortest Path Problem Fangguo He Huan Qi Qiong Fan Institute of Systems Engineering, Huazhong University of Science & Technology, Wuhan 430074, P. R. China

More information

Crosstalk-Aware Spectrum Defragmentation based on Spectrum Compactness in SDM-EON

Crosstalk-Aware Spectrum Defragmentation based on Spectrum Compactness in SDM-EON Crosstalk-Aware Spectrum Defragmentation based on Spectrum Compactness in SDM-EON Yongli Zhao State Key Laboratory of Information Photonics and Optical Communications, Beijing University of Posts and Telecommunications

More information

Deadline-Aware Co-Scheduling Using Anycast Advance Reservations in Wavelength Routed Lambda Grids

Deadline-Aware Co-Scheduling Using Anycast Advance Reservations in Wavelength Routed Lambda Grids Deadline-Aware Co-Scheduling Using Anycast Advance Reservations in Wavelength Routed Lambda Grids Hitesh Kulkarni, Arush Gadkar, and Vinod M. Vokkarane Department of Computer and Information Science University

More information

Machine-Learning-Based Flow scheduling in OTSSenabled

Machine-Learning-Based Flow scheduling in OTSSenabled Machine-Learning-Based Flow scheduling in OTSSenabled Datacenters Speaker: Lin Wang Research Advisor: Biswanath Mukherjee Motivation Traffic demand increasing in datacenter networks Cloud-service, parallel-computing,

More information

OPTICAL NETWORKS. Optical Metro Networks. A. Gençata İTÜ, Dept. Computer Engineering 2005

OPTICAL NETWORKS. Optical Metro Networks. A. Gençata İTÜ, Dept. Computer Engineering 2005 OPTICAL NETWORKS Optical Metro Networks A. Gençata İTÜ, Dept. Computer Engineering 2005 Introduction Telecommunications networks are normally segmented in a three-tier hierarchy: Access, metropolitan,

More information

Iterative Optimization in VTD to Maximize the Open Capacity of WDM Networks

Iterative Optimization in VTD to Maximize the Open Capacity of WDM Networks Iterative Optimization in VTD to Maximize the Open Capacity of WDM Networks Karcius D.R. Assis, Marcio S. Savasini and Helio Waldman DECOM/FEEC/UNICAMP, CP. 6101, 13083-970 Campinas, SP-BRAZIL Tel: +55-19-37883793,

More information

Design of Hierarchical Crossconnect WDM Networks Employing a Two-Stage Multiplexing Scheme of Waveband and Wavelength

Design of Hierarchical Crossconnect WDM Networks Employing a Two-Stage Multiplexing Scheme of Waveband and Wavelength 166 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 20, NO. 1, JANUARY 2002 Design of Hierarchical Crossconnect WDM Networks Employing a Two-Stage Multiplexing Scheme of Waveband and Wavelength

More information

Traffic Grooming and Regenerator Placement in Impairment-Aware Optical WDM Networks

Traffic Grooming and Regenerator Placement in Impairment-Aware Optical WDM Networks Traffic Grooming and Regenerator Placement in Impairment-Aware Optical WDM Networks Ankitkumar N. Patel, Chengyi Gao, and Jason P. Jue Erik Jonsson School of Engineering and Computer Science The University

More information

Dynamic Wavelength Assignment for WDM All-Optical Tree Networks

Dynamic Wavelength Assignment for WDM All-Optical Tree Networks Dynamic Wavelength Assignment for WDM All-Optical Tree Networks Poompat Saengudomlert, Eytan H. Modiano, and Robert G. Gallager Laboratory for Information and Decision Systems Massachusetts Institute of

More information

Adapted and Constrained Dijkstra for Elastic Optical Networks

Adapted and Constrained Dijkstra for Elastic Optical Networks Adapted and Constrained Dijkstra for Elastic Optical Networks Ireneusz Szcześniak AGH University of Science and Technology Department of Telecommunications al. Mickiewicza 30 30-059 Krakow Poland Bożena

More information

A Customizable Two-Step Framework for General Equipment Provisioning in Optical Transport Networks

A Customizable Two-Step Framework for General Equipment Provisioning in Optical Transport Networks A Customizable Two-Step Framework for General Equipment Provisioning in Optical Transport Networks Limin Tang, Wanjun Huang, Miguel Razo, Arularasi Sivasankaran, Paolo Monti, Marco Tacca, Andrea Fumagalli

More information

Wavelength-Switched to Flex-Grid Optical Networks

Wavelength-Switched to Flex-Grid Optical Networks Book Chapter Review-Evolution from Wavelength-Switched to Flex-Grid Optical Networks Tanjila Ahmed Agenda ØObjective ØIdentifying the Problem ØSolution: Flex-Grid Network ØFixed-grid DWDM Architecture

More information

MODERN communication networks are constructed

MODERN communication networks are constructed 1000 IEEE/ACM TRANSACTIONS ON NETWORKING, VOL. 19, NO. 4, AUGUST 2011 Cross-Layer Survivability in WDM-Based Networks Kayi Lee, Member, IEEE, Eytan Modiano, Senior Member, IEEE, and Hyang-Won Lee, Member,

More information

A Novel High Performance Multicast Scheme on Virtual Ring-Based 2D Torus Topology in DWDM Networks

A Novel High Performance Multicast Scheme on Virtual Ring-Based 2D Torus Topology in DWDM Networks Tamkang Journal of Science and Engineering, Vol. 14, No. 1, pp. 81 89 (2011) 81 A Novel High Performance Multicast Scheme on Virtual Ring-Based 2D Torus Topology in DWDM Networks I-Shyan Hwang 1 *, San-Nan

More information

CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION

CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION 131 CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION 6.1 INTRODUCTION The Orthogonal arrays are helpful in guiding the heuristic algorithms to obtain a good solution when applied to NP-hard problems. This

More information

Capacity planning and.

Capacity planning and. Some economical principles Hints on capacity planning (and other approaches) Andrea Bianco Telecommunication Network Group firstname.lastname@polito.it http://www.telematica.polito.it/ Assume users have

More information

Spectrum Assignment in Optical Networks: A Multiprocessor Scheduling Perspective

Spectrum Assignment in Optical Networks: A Multiprocessor Scheduling Perspective 754 J. OPT. COMMUN. NETW./VOL. 6, NO. 8/AUGUST 204 Talebi et al. Spectrum Assignment in Optical Networks: A Multiprocessor Scheduling Perspective Sahar Talebi, Evripidis Bampis, Giorgio Lucarelli, Iyad

More information

An Efficient Algorithm for Virtual-Wavelength-Path Routing Minimizing Average Number of Hops

An Efficient Algorithm for Virtual-Wavelength-Path Routing Minimizing Average Number of Hops An Efficient Algorithm for Virtual-Wavelength-Path Routing Minimizing Average Number of Hops Harsha V. Madhyastha Department of Computer Science and Engineering Indian Institute of Technology, Madras Chennai,

More information

Performance Evaluation of Brute Force Techniques for Routing and Spectrum Assignment in Elastic Optical Network using MPI and CUDA

Performance Evaluation of Brute Force Techniques for Routing and Spectrum Assignment in Elastic Optical Network using MPI and CUDA Performance Evaluation of rute Force Techniques for Routing and Spectrum Assignment in Elastic Optical Network using MPI and CUDA Mahmoud Fayez mahmoud.fayez@cis.asu.edu.eg Tarek F. Gharib Information

More information

Splitter Placement in All-Optical WDM Networks

Splitter Placement in All-Optical WDM Networks plitter Placement in All-Optical WDM Networks Hwa-Chun Lin Department of Computer cience National Tsing Hua University Hsinchu 3003, TAIWAN heng-wei Wang Institute of Communications Engineering National

More information

Some economical principles

Some economical principles Hints on capacity planning (and other approaches) Andrea Bianco Telecommunication Network Group firstname.lastname@polito.it http://www.telematica.polito.it/ Some economical principles Assume users have

More information