Lightpath Protection using Genetic Algorithm through Topology Mapping in WDM Optical Networks

Utpal Biswas†, Anirban Mukhopadhyay†, Ujjwal Maulik+, and Mrinal Kanti Naskar* †

Department of Computer Science and Engineering, University of Kalyani, Kalyani – 741235, India

+

*

Department of Computer Science and Engineering, Jadavpur University, Kolkata – 700032, India

Department of Electronics and Telecommunication Engineering, Jadavpur University, Kolkata – 700032, India Email: [email protected], [email protected], [email protected], [email protected]

Abstract In WDM optical networks, design protection algorithms aim to maintain the connectivity between two nodes in the network following a failure by mapping a given virtual topology onto the physical topology so as to minimize the failure propagations in a network. In this article, a genetic algorithm based approach has been proposed for design protection in WDM optical networks. The proposed method has been compared with two other well-known design protection algorithms to establish its effectiveness.

Keywords: WDM optical networks, lightpath protection, genetic algorithm, topology mapping

1. Introduction

Fiber optic technology is emerging as the most promising candidate for the next generation bandwidth intensive applications and high speed computer networks because of huge bandwidth of optical fibers and capability of multiplexing channels onto a single fiber through wavelength division multiplexing (WDM) technique. A pointto-point interconnection of optical wavelength routers using optical links is known as physical topology of a WDM optical network. A virtual (or logical) topology of a network is a graph consisting of all the lightpaths present in the network. A lightpath between two nodes is defined by a sequence of links between the nodes and a wavelength on these links. Same wavelength must be used on all the links in the sequence and this is known as

1

wavelength continuity constraint [1]. As a result, two lightpaths sharing a common link cannot use the same wavelength on their paths.

Survivability refers to the ability of a network to reconfigure and reestablish communication upon failures. It is an important issue to be addressed at the time of designing a network. The issue becomes more relevant in WDM optical networks, as a single link failure may result into failure of several wavelength channels or lightpaths. Hence, extra care must be taken to find out the ways to protect these lightpaths against failures. Survivability in a network can be provided either through protection or restoration. Protection is the primary mechanism used to deal with a failure and need to be very fast. In contrast, restoration is used to provide more efficient routes or additional resilience against further failures. Protection in WDM can be performed at two levels, namely physical and design. In physical protection, a failure situation can be handled by optical hardware duplication without the higher level switches being aware of the failure and hence the scheme may not be cost effective. On the other hand, design protection aims at maintaining connectivity between two nodes in the network following a failure by mapping a given virtual topology onto the physical topology so as to minimize the failure propagations in a network. In reference [3], a brief survey of protection / restoration schemes can be found. As far as our knowledge goes, none of the earlier schemes have used genetic algorithm to solve the design protection problem.

In this paper, a genetic algorithm for topology mapping to protect lightpaths in a WDM optical network has been proposed. The paper is organized as follows. Section 2 illustrates the problem statement and its mathematical formulations. In section 3, a genetic algorithm for solving the problem is presented. In section 4, the performance of the genetic algorithm is compared with that of other two algorithms. Section 5 finally concludes the paper.

2. Design Protection Problem

In this section, we formally define the design protection as an optimization problem [2] and give its mathematical formulation. 2.1. Problem Statement

2

Given the physical topology of a network and a virtual topology to be mapped onto the physical topology, the problem is to find a route for each of the lightpaths in the network so that the number of broken lightpaths due to a link failure is minimized.

2.2. Mathematical Formulation

The following notation is used:

N

: total number of nodes in the network.

s, d : source and destination nodes of a lightpath, respectively.

a, b : end points of a link.

Given

i) E = (eab) , an N X N symmetric matrix representing the physical topology, where,

w, the number of wavelength channels on a link. eab = 0, if there is no link between a and b.

ii) L = (lsd), an N X N symmetric matrix representing the virtual topology, where,

1, if there is a lightpath between s and d. lsd = 0, otherwise.

3

Variables

i) Psd = ( pabsd), an N X N symmetric matrix representing the route of the lightpath lsd on the physical topology, where,

1, if lsd uses link eab on its route. pabsd = 0, otherwise.

ii) L′ab = ( l′sd ), an N X N symmetric matrix representing the virtual topology after removing all the lightpaths that have eab as part of their route.

0, if (lsd = 0) ∪ (pabsd =1). l′sd = 1, otherwise.

iii) Kabsd = 1, if lightpath lsd has no alternate path when link eab is broken.

1, if (lsd =1) ∩ (eab > 0) ∩ (s and d unconnected in L′ab ). Kabsd = 0, otherwise.

A broken pair is defined as a tuple (lsd, eab) for which Kabsd = 1.

Objective function

Minimize

∑∑ K

sd ab

subject to the following constraints:

a ,b s , d

4

i) Lightpaths do not exceed the maximum capacity of each link, i.e.,

∑p

sd ab

≤ w , for all a, b.

s ,d

ii) A lightpath does not encounter a node twice in its route, i.e.,

∑p

sd ab

≤2

a

3. Design Protection Using Genetic Algorithm

A genetic algorithm (GA) [4] is mainly a probabilistic search algorithm based on the principles and concept of natural selection and evolution. At each generation it maintains a population of individuals where each individual is a coded form of a possible solution of the problem at hand and called chromosome. Each chromosome is evaluated by a function known as fitness function which is usually the inverse of the objective function of the corresponding optimization problem. Next, new population is generated from the present one through selection, crossover and mutation operations. Purpose of selection mechanism is to select more fit individuals (parents) for crossover and mutation. Crossover causes the exchange of genetic materials between the parents to form offspring, whereas mutation incorporates new genetic material in the offspring. Implementations of above mentioned components for the proposed genetic algorithm are as follows.

3.1. Chromosome Representation

Each chromosome is represented by an integer string of length N X L, where N is the number of nodes in the physical network and L is the number of lightpaths in the virtual topology.

Fig 1. Chromosome representation

5

In this representation, each chromosome encodes a network configuration as one of the solutions to the problem and each chromosome segment represents a lightpath. An example illustrating chromosome segment generation is shown below in Table1 for the network configuration 1 in Fig 3.

Table1: Generation of chromosome segments Lightpath 0: Shortest path between nodes 0 and 1.

0

3

1

-

-

-

-

-

Chrom1_segS1

Lightpath 1: Shortest path between nodes 0 and 5.

0

3

5

-

-

-

-

-

Chrom1_segS2

Lightpath 2: Shortest path between nodes 1 and 4.

1

3

5

4

-

-

-

-

Chrom1_segS3

Lightpath 3: Shortest path between nodes 2 and 3.

2

1

3

-

-

-

-

-

Chrom1_segS4

Lightpath 4: Shortest path between nodes 2 and 5.

2

1

3

5

-

-

-

-

Chrom1_segS5

The overall chromosome configuration is represented by cascading the above mentioned chromosome segments.

3.2. Initial Population

Initial population is generated by selecting random paths between two nodes connected by a lightpath. It is ensured that the random paths do not contain a node twice i.e. no cycle is included in the path.

3.3. Fitness Function

In our work, the fitness function is taken as the reciprocal of the total number of lightpaths damaged due to link failures in a particular network configuration.

3.4. Selection Operation

The probability of selection of a network configuration is directly proportional to the fitness value of the corresponding chromosome. Hence, a configuration with greater fitness value has a greater chance of being selected for crossover and mutation. In this work, we have used ‘Roulette wheel’ [4] as the selection mechanism.

6

3.5. Crossover Operation

The crossover strategy used in this implementation is that of number crossover applied separately on each lightpath segment of a chromosome. Let, chrom1_segS and chrom2_segS are two corresponding lightpath segments from two different chromosomes on which crossover is performed. End nodes of both the segments must be same as they correspond to the same lightpath. A random crossover point p1 corresponding to chrom1_segS is generated between 0 and (k–1), where length of random path in chrom1_segS is k. Similarly, a random crossover point p2, corresponding to chrom2_segS is generated between 0 and (l–1), where the length of random path in chrom2_segS is l. New chromosome segments (offspring segments) are formed by copying chrom1_segS from 0 to p1 and chrom2_segS from p2 to (l–1) in offspring1_segS and copying chrom2_segS from 0 to p2 and chrom1_segS from p1 to (k–1) in offspring2_segS. After crossover, it is possible to have repetition of a node in the offspring segments.

Chrom1_segS

0

p1 k

N

Chrom2_segS

0 p2

l

N

offspring1_segS

offspring2_segS

0

0 p2

p1

N

N

Fig 2. Illustration of crossover operation

3.6. Mutation Operation

Mutation is performed by selecting an alternate path between two randomly selected nodes of a chromosome segment. The alternate path between two random nodes is generated by applying the shortest path algorithm on the network obtained after the deletion of the first shortest path between the nodes.

A new solution obtained through crossover and mutation is accepted only if its fitness value is better than that of the worst solution in the population. Otherwise, the new solution is rejected.

7

4. Results

and Discussions

In order to determine the effectiveness of the proposed GA based protection algorithm, its performance has been compared with that of the two other algorithms, viz., Shortest path routing algorithm (SPRA) [2] and Heuristic algorithm (HA) [5].

All the algorithms have been implemented on Pentium machine using C as the programming language. Experiments are conducted on a number of different network configurations, whereas, the results on two of them as shown in Fig. 3 and Fig. 4, respectively, are reported here. For each network configuration, the number of wavelength channels permitted is assumed to be 5. Moreover, the following parameters are assumed for the GA: Probability of crossover = 0.95, Probability of mutation = 0.03, Number of generations = 100, Population size = 4.

Fig 3. Network configuration 1

Fig 4. Network configuration 2

8

Table 2 reports the number of broken pairs achieved by the three algorithms for both the network configurations. For Network 1, GA achieves 7 broken pairs which is better than both SPRA (12 broken pairs) and HA (8 broken pairs). For Network 2, both HA and GA provide 2 broken pairs, whereas SPRA gives 10 broken pairs. Table 2: Comparative results of different algorithms Algorithm

Number of Broken Pairs Network 1

Network 2

SPRA

12

10

HA

8

2

GA

7

2

As the network complexity increases with large number of nodes, physical links and lightpaths the superiority of the genetic algorithm becomes evident. Moreover, the performance of the genetic algorithm may further be improved by increasing the number of generations.

5. Conclusions

In this work, a genetic algorithm for design protection problem in optical networks, employing wavelengthdivision multiplexing (WDM) is proposed. Our main objective was to map a given virtual topology onto a physical topology so as to minimize the failure propagations in the network. A comparative study of the proposed genetic algorithm with two other algorithms found in the literature, has been carried out. The results of our study show that the genetic algorithm gives better solutions than the others.

References

[1] A Survey of Virtual Topology Design Algorithms for wavelength routed Optical Networks: Rudra Dutta and George N. Rouskas, Optical Networks Magazine, 73 (Jan 2000). [2] Design Protection for WDM Optical Networks: Olivier Crochat and J. Y. L. Boudec, IEEE JSAC, Vol 16, 1158 (1998).

9

[3] Restorations Schemes in WDM Optical Networks: M. K. Naskar, U. Biswas, S. K. Sarkar, Nat.Conf. NCSSS, Coimbatore (March 2002). [4] Genetic Algorithms in search, Optimization & Machine Learning: D. E. Goldberg, A Gentle Introduction to Genetic Algorithms, Addison Wesley, Singapore, (1989). [5] An Approach to Wide Area WDM Optical Network Design using Genetic Algorithm: D. Saha, M. D. Purkayastha, A. Mukherjee, Computer Communications, 22, 156 (1999).

10

Lightpath Protection using Genetic Algorithm ... - Semantic Scholar

connectivity between two nodes in the network following a failure by mapping ... applications and high speed computer networks because of huge bandwidth of ...

60KB Sizes 1 Downloads 437 Views

Recommend Documents

Lightpath Protection using Genetic Algorithm ... - Semantic Scholar
virtual topology onto the physical topology so as to minimize the failure ... applications and high speed computer networks because of huge bandwidth of optical ...

Lightpath Protection using Genetic Algorithm through ...
... Jadavpur University, Kolkata – 700032, India. *Department of Electronics and Telecommunication Engineering, Jadavpur University, Kolkata – 700032, India.

1 feature subset selection using a genetic algorithm - Semantic Scholar
Department of Computer Science. 226 Atanaso Hall. Iowa State ...... He holds a B.S. in Computer Science from Sogang University (Seoul, Korea), and an M.S. in ...

a niche based genetic algorithm for image ... - Semantic Scholar
Image registration can be regarded as an optimization problem, where the goal is to maximize a ... genetic algorithms can address this problem. However ..... This is akin to computing the cosine ... Also partial occlusions (e.g. clouds) can occur ...

Dynamic Channel Allocation Using a Genetic ... - Semantic Scholar
categorised in terms of a Channel Allocation Matrix. A novel. Dynamic ... same category in the proposed Channel Allocation Matrix. .... Priority decrease. 3 5 2 11 9. Fig. 5. An example of an individual's string representation. The fitness value is e

Dynamic Channel Allocation Using a Genetic ... - Semantic Scholar
Internet users that do not enjoy fast access networks. Broadband fixed wireless access (BFWA) networks can be rapidly deployed and provide data rates that current Internet users demand. Typical BFWA network components are the ..... operating in the 5

Eye Movement Data Modeling Using a Genetic ... - Semantic Scholar
Yun Zhang is with School of Computer Science, Northwestern. Polytechnical University ... Manu. Tech./ Northwestern Polytechnical University, Xi'an, Shaanxi, P.R. .... movements that reach a velocity of over 500 degrees/second. From less ...

Genetic diversity in palmyrah genotypes using ... - Semantic Scholar
Email: [email protected] collection of neera and .... amplified the products in all the templates used. The ... DNA as template, 1 unit(0.3 µl) of Taq DNA.

Eye Movement Data Modeling Using a Genetic ... - Semantic Scholar
Dagan Feng is also with School of Information Technologies, The. University of .... effective receptor density falls dramatically within one degree of the central ...

Oxidative Preconditioning Affords Protection ... - Semantic Scholar
dard laboratory chow and water ad libitum and were kept ..... is noteworthy that plants also can express a protec- ... Port City Press, Inc.: New York, 1993;.

Modified Aho Corasick Algorithm - Semantic Scholar
apply the text string as input to the pattern matching machine. ... Replacing this algorithm with the finite state approach resulted in a program whose running.

Modified Aho Corasick Algorithm - Semantic Scholar
goto function g(0,a)=s, where s is valid if a={h,s}.then this function makes a transition from State 0 to state 1 or 3 depending on the symbol. In the current example ...

using rapd markers - Semantic Scholar
based on this, cluster analysis was done using minimum variance algorithm. Cluster analysis showed two major groups. Each sub-group was characterized ...

using rapd markers - Semantic Scholar
RAPD data were used to calculate a Squared Euclidean Distance matrix, and based on this, cluster ... Africa, South-East, Asia, U.S.A, Brazil, Australia and. Turkey. In some ... homogenate was cooled to room temperature and extracted with 5 ...

Event Detection in Baseball Videos Using Genetic Algorithm ... - APSIPA
Department of Computer Science and Information Engineering, National Chiayi University ..... consider that the BS measure is formulated based on SIFT-.

Using genetic algorithm to identify the discriminatory ...
classes in the category layer are represented by a bi- nary string with a value ..... testing) using MLP-BP is less than the FA due to the simpler architecture of the.