Towards realistic physical topology models for Internet backbone networks

In this paper 1, we consider the problem of physical topology design (i.e., physical connectivity) for Internet backbone networks. We explore the driving forces for service providers to layout fiber links, and propose a new problem formulation that can accurately emulate the existing optical backbone networks. Unlike previous studies which mainly focused on deployment cost, our model captures the physical design principles including (1) the cost of the infrastructure, (2) the expected performance, (3) geographical constraints, and (4) the resilience of the network to link/node failures (survivability). Obtaining an optimal solution is shown to be NP-hard, we thus present a polynomial time heuristic algorithm, HINT, to determine the number and the choice of constituent links. The efficacy of HINT is established in comparison with the published maps of three major scientific and commercial backbone networks: Internet2 Abilene, AT&T domestic express backbone, and Level3 network. Preliminary results reveal that taking performance, resilience and geographical constraints into consideration is necessary to emulate real backbones. The HINT heuristic yields a similarity of more than 90% with the published structures.

[1]  Albert G. Greenberg,et al.  Experience in measuring backbone traffic variability: models, metrics, measurements and meaning , 2002, IMW '02.

[2]  Biswanath Mukherjee,et al.  Some principles for designing a wide-area WDM optical network , 1996, TNET.

[3]  T. C. Hu.  Optimum Communication Spanning Trees , 1974, SIAM J. Comput..

[4]  D. Habibi,et al.  Establishing Physical Survivability of Large Networks using Properties of Two-Connected Graphs , 2005, TENCON 2005 - 2005 IEEE Region 10 Conference.

[5]  Eytan Modiano,et al.  Survivable lightpath routing: a new approach to the design of WDM-based networks , 2002, IEEE J. Sel. Areas Commun..

[6]  Chi Guan,et al.  Efficient physical topologies for regular WDM networks , 2004, Optical Fiber Communication Conference, 2004. OFC 2004.

[7]  Gwendolyn Halford,et al.  County and City Data Book , 1999 .

[8]  Walter Willinger,et al.  A first-principles approach to understanding the internet's router-level topology , 2004, SIGCOMM '04.

[9]  Walter Willinger,et al.  Understanding Internet topology: principles, models, and validation , 2005, IEEE/ACM Transactions on Networking.

[10]  George N. Rouskas,et al.  On the physical and logical topology design of large-scale optical networks , 2003 .

[11]  Kumar N. Sivarajan,et al.  Design of Logical Topologies for Wavelength-Routed Optical Networks , 1996, IEEE J. Sel. Areas Commun..

[12]  Huan Liu,et al.  Physical Topology Design for All-Optical Networks , 2006, 2006 3rd International Conference on Broadband Communications, Networks and Systems.

[13]  M. C. Sinclair,et al.  Design and dimensioning of dual-homing hierarchical multi-ring networks , 2000 .

[14]  Biswanath Mukherjee,et al.  Wavelength-routed optical networks: linear formulation, resource budgeting tradeoffs, and a reconfiguration study , 2000, TNET.

[15]  A. Dwivedi,et al.  Traffic model for USA long-distance optical network , 2000, Optical Fiber Communication Conference. Technical Digest Postconference Edition. Trends in Optics and Photonics Vol.37 (IEEE Cat. No. 00CH37079).

[16]  Gaoxi Xiao,et al.  Two-stage cut saturation algorithm for designing all-optical networks , 2001, IEEE Trans. Commun..

[17]  Chen-Nee Chuah,et al.  Feasibility of IP restoration in a tier 1 backbone , 2004, IEEE Network.

[18]  Ratul Mahajan,et al.  Measuring ISP topologies with Rocketfuel , 2004, IEEE/ACM Transactions on Networking.

[19]  Hawoong Jeong,et al.  Modeling the Internet's large-scale topology , 2001, Proceedings of the National Academy of Sciences of the United States of America.

[20]  Kathryn Fraughnaugh,et al.  Introduction to graph theory , 1973, Mathematical Gazette.