Heuristic algorithms for efficient allocation of multicast-capable nodes in sparse-splitting optical networks☆
Introduction
Optical networks have evolved steadily over the last two decades from wavelength division multiplexed (WDM) point-to-point systems at the physical layer providing transport capabilities through optical fibers, to ring, and subsequently mesh topologies with intelligent switching elements (reconfigurable optical add-drop multiplexers (ROADMs), optical cross-connects (OXCs), etc.) that can now provide provisioning of wavelength and sub-rate connections, fault accommodation, as well as several other control functionalities at the physical (optical) layer. With the successful commercialization of WDM, and several key technology advancements of optical component technologies (such as optical amplifiers, lasers, filters, and optical switches amongst others) within the optical networking space, the standardized optical transport network (OTN) nowadays provides for carrier-grade operations, administration, and maintenance (OAM) for managed wavelength services, as well as fault accommodation for high service availability [1].
Next-generation optical networks are expected to support traffic that will be heterogeneous in nature with both unicast, as well as multicast applications. Even though most connections carried over an optical mesh network are still currently unicast connections (e.g., high-bandwidth point-to-point connections for enterprise customers) new traffic requirements and applications are driving the evolution of the network architectures, requiring multicast capabilities to deliver high-bandwidth content. For example, recent bandwidth-intensive applications that are driving the use of optical multicasting include telepresence, grid computing, telemedicine, software and video distribution for residential customers, movie broadcasts, interactive distance learning and video training, and distributed games amongst others.
Multicasting refers to the simultaneous transmission of information from a single source to several destinations. Optical multicast requests are established via the provisioning of trees (called light-trees in optical networks), that are created utilizing optical splitters at the network nodes [2], [3]. Thus, in order to support these multicast connections, the utilization of multicast-capable nodes (nodes where optical splitting can take place), strategically placed at certain node locations during the network design phase, is of great interest, as it will provide efficient multicast connectivity while keeping the network cost low (by not utilizing MC nodes throughout the entire network). This results in a sparse-splitting network [2], [3], where some of the network nodes are multicast-capable, while the rest are multicast-incapable (MI) (nodes that do not have optical splitting capabilities). These MI nodes can also be distinguished as Drop-and-Continue (DaC) or Drop-or-Continue (DoC) nodes. A DaC node can transmit the optical signal to the following node in its path and can also drop it locally as well, while a DoC node can either transmit the optical signal to the following node in its path or drop it locally. Since both networks architectures are viable possibilities [4], [5], the current paper deals with both DaC and DoC networks. The analysis of both cases can subsequently be utilized by network engineers and designers to ascertain both architectures when deciding what technologies and architectures to deploy in their networks.
As the problem of where to optimally place the MC nodes in the network (MC node allocation) is an NP-complete problem [6], polynomial-time heuristics that give approximate solutions are used in practice. This is precisely the focus of this work. In the current paper three heuristics are proposed for efficient MC node allocation, that can be applied for both DoC and DaC networks. Their performance evaluation, through simulations on the well-known USNET and NSFNET networks as well as on larger, randomly created networks, has shown that they achieve an important decrease of the average cost of the derived multicast trees compared to the conventional placement methods. Furthermore, this work also investigates the impact of networks having DaC rather than DoC capabilities, as well as the impact of the percentage of MC nodes on the network performance, providing guidance for the efficient design of optical networks with sparse multicasting capabilities.
The remaining of the paper is organized as follows: The problem formulation is given in Section 2, as well as the notation and definitions that are used throughout the paper. The existing work on MC node allocation is presented in Section 3. Section 4 presents the proposed techniques for cost-efficient allocation of MC nodes, while their performance evaluation is presented in Section 5. Finally, in Section 6, the conclusions of the paper are presented, as well as directions for future work.
Section snippets
Problem formulation, notation, definitions
Throughout the paper, the following notation and definitions are utilized.
- •
The network is modeled as a directed graph where V () and A () are the sets consisting of the network nodes (representing the optical switching nodes) and arcs (representing the optical fibers), respectively.
- •
The notation [i, j] stands for the arc originating from node i and ending at node j.
- •
A cost cij is assigned to each arc [i, j].
- •
The network directed graph is considered to be symmetric: for every arc
Existing work
Existing methods for allocating the limited number of MC nodes into the network can be found in [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16] as well as other sources. In [9], [10] the techniques of k-Maximum Degree (kmaxD) and k-Maximum Wavelength Reduction (kmaxWR) are presented, as explained below.
k-Maximum Degree (kmaxD) method: The key idea of this approach is that a node with more neighboring nodes is more likely to become a branch node of a multicast tree. Hence, placing a
Proposed heuristics
Three heuristics for cost-efficient allocation of MC nodes are proposed, namely Decreased Number of Branches Heuristic (DNB), Least Useful Removed First Heuristic (LURF), and Most Useful Added First Heuristic (MUAF) as described below.
Evaluation on the USNET and NSFNET Networks
The performance of the proposed heuristic algorithms for efficient allocation of the limited number of MC nodes was evaluated through simulations on the widely used USNET network [22], consisting of 24 nodes and 43 links (Fig. 3) as well as on the also widely used NSFNET network [23], consisting of 14 nodes and 22 links (Fig. 4). In both networks each link consists of a pair of arcs of opposite orientation. A cost is assigned to each network link as shown in Figs. 3 and 4, and multiple
Conclusions
In the current paper, three heuristics were presented for efficient allocation of the limited number of multicast-capable nodes in sparse-splitting optical networks. Both cases of DoC and DaC networks were investigated. Simulations on the well-known USNET and NSFNET networks as well as on larger, randomly created networks, have shown that they achieve an important decrease of the average cost of the derived multicast trees compared to the relevant conventional methods. Furthermore, this work
Costas K. Constantinou holds a Bachelor in Physics from the Aristotelian University of Thessaloniki and a Ph.D. in Electrical Engineering from the University of Cyprus. He is a Research Fellow at the KIOS Research Center for Intelligent Systems and Networks, University of Cyprus. His research interests focus on the areas of optical networks, transportation networks, routing algorithms and graph theory.
References (30)
- et al.
Placement of wavelength converters and light splitters in a WDM network using the generic graph model
Comput. Commun.
(2010) - et al.
Sparse-partial wavelength conversion: converter placement and wavelength assignment
Opt. Switch. Netw.
(2010) - et al.
Strategies and techniques for node placement in wireless sensor networks: a survey
Ad Hoc Netw.
(2008) - et al.
Multiwavelength Optical Networks: Architectures, Design, and Control
(2008) - et al.
Light-trees: optical multicasting for improved performance in wavelength-routed networks
IEEE Commun. Mag.
(1999) Optical Communication Networks
(1997)- et al.
Constrained multicast routing in WDM networks with sparse light splitting
IEEE/OSA J. Lightwave Technol.
(2000) - et al.
All-optical multicasting on wavelength-routed WDM networks with partial replication
Proceedings of International Conference on Information Networking, Beppu City, Oita, Japan
(Jan.-Feb. 2001) - et al.
Allocation of splitting nodes in all-optical wavelength-routed networks
Photonic Netw. Commun.
(2000) - et al.
Designing and engineering metropolitan area transparent optical networks for the provisioning of multicast sessions
Proceedings of the IEEE/OSA Optical Fiber Communications Conference (OFC), San Diego, CA
(March 2010)
Multicast routing algorithms based on Q-factor physical-layer constraints in metro networks
IEEE Photonics Technol. Lett.
Splitter placement in all-optical WDM networks
Proceedings of IEEE Global Telecommunications Conference (GLOBECOM)
Allocation of light splitters in all-optical WDM networks with sparse light splitting capabilities
Telecommun. Syst.
Allocation of multicast nodes in wavelength-routed networks
Proceedings of IEEE International Conference on Communications (ICC)
Optimization of splitting node placement in wavelength-routed optical networks
IEEE J. Sel. Areas Commun.
Cited by (6)
Lotka-Volterra distributed power control model for OCDMA systems
2021, AEU - International Journal of Electronics and CommunicationsCitation Excerpt :Herein, the analytical solutions are explored through continuous optimization methods, namely sequential quadratic programming (SQP) and augmented Lagrangian method (ALM). It is worth noting that in the literature, the optimization approach commonly applied to solve power assignment problems in optical networks has been heuristic methods, for instance [9,10,11,12]. In [13] for CDMA networks, the authors use an iterative analytical method using the Verhuslt population model to represent a new DPCA.
Hierarchical Structure and Placement of MC Node in Optical WDM Network
2023, Lecture Notes in Networks and SystemsEfficient Placement of Splitters in Optical WDM Network
2021, Research SquareMulticast routing and allocation of wavelengths in a WDM network with splitters and converters
2021, International Journal of Internet Protocol TechnologyJoint Optimization for Multicast Provisioning in Mixed-Line-Rate Optical Networks with a Column Generation Approach
2018, Journal of Lightwave TechnologyOptimal and heuristic algorithms for all-optical group multicast in resource-constrained WDM networks
2016, International Journal of Communication Systems
Costas K. Constantinou holds a Bachelor in Physics from the Aristotelian University of Thessaloniki and a Ph.D. in Electrical Engineering from the University of Cyprus. He is a Research Fellow at the KIOS Research Center for Intelligent Systems and Networks, University of Cyprus. His research interests focus on the areas of optical networks, transportation networks, routing algorithms and graph theory.
Georgios Ellinas holds a B.S. (1991), M.Sc. (1993), M.Phil. (1995), and a Ph.D. (1998) in Electrical Engineering from Columbia University. Dr. Ellinas is currently an Associate Professor and the Chair of the Department of Electrical and Computer Engineering at the University of Cyprus. Prior to joining the University of Cyprus Dr. Ellinas was an Associate Professor of Electrical Engineering at City College of the City University of New York (2002–2005). Before joining the academia, Dr. Ellinas was a Senior Network Architect at Tellium Inc (2000–2002). Dr. Ellinas also served as a Visiting Scientist/Research Scientist in Telcordia Technologies’ (formerly Bellcore) Optical Networking Research Group (1993–2000), and as an Adjunct Assistant Professor at Columbia University and the University of Maryland Baltimore County in 1999 and 2000, respectively. He has co-authored two books on optical networks (Wiley 2007, Cambridge University Press 2008), he is the co-editor of another book on optical networks (Springer 2011), he has authored/co-authored more than 190 journal and conference papers and book chapters, and he is the holder of 30 patents on optical networking. His research interests focus on optical and converged optical-wireless networks, intelligent transportation systems, critical infrastructure systems, and the Internet of Things. He is a Senior Member of IEEE, and a Member of OSA, ACM, and the Marie Curie Fellows Association.
- ☆
This work was supported by the Cyprus Research Promotion Foundation’s Framework Programme for Research, Technological Development and Innovation 2009 (DESMI 2009–2010), co-funded by the Republic of Cyprus and the European Regional Development Fund, and specifically under Grant TPE/EPIKOI/0311(BIE)/11.