ABSTRACT
Network design is the process of dimensioning IP capacity over an optical network infrastructure to satisfy a given set of demands and reliability constraints. Specifically, we consider the problem of hose-based cross-layer network design, which seeks to find a minimum cost design that is able to route demand for all hose traffic matrices under all specified failure states. While most network design problems are solved as Mixed Integer Programs, a commercial solver can become intractable due to the scale of today's networks. We demonstrate how the classic Benders decomposition algorithm can be applied and improved for this problem and discuss practical implementation aspects. We showcase a horizontally scalable distributed framework to leverage the decomposable problem structure and solve millions of linear programs in a distributed manner, thereby making the network design problem tractable. In contrast to the conventional approach where failure states and traffic matrices are planned sequentially, the Benders algorithm finds global optimal designs across all traffic matrices and failure states. This leads to network designs with improved solution quality and reliability, with 20--30% less IP capacity and spectrum consumption, 50% less link augments and up to 20x faster runtime that enables design for hyper scale networks in a matter of hours.
- Satyajeet Singh Ahuja, Vinayak Dangui, Kirtesh Patil, Manikandan Somasundaram, Varun Gupta, Mario Sanchez, Guanqing Yan, Max Noormohammadpour, Alaleh Razmjoo, Grace Smith, Hao Zhong, Abhinav Triguna, Soshant Bali, Yuxiang Xiang, Yilun Chen, Prabhakaran Ganesan, Mikel Jimenez Fernandez, Petr Lapukhov, Guyue Liu, and Ying Zhang. 2022. Network entitlement: contract-based network sharing with agility and SLO guarantees. In Proceedings of the ACM SIGCOMM 2022 Conference. 250--263.Google ScholarDigital Library
- Satyajeet Singh Ahuja, Varun Gupta, Vinayak Dangui, Soshant Bali, Abishek Gopalan, Hao Zhong, Petr Lapukhov, Yiting Xia, and Ying Zhang. 2021. Capacity-efficient and uncertainty-resilient backbone network planning with hose. In Proceedings of the ACM SIGCOMM 2021 Conference. 547--559.Google ScholarDigital Library
- Satyajeet Singh Ahuja, Petr Lapukhov, Varun Gupta, Vinayak Dangui, Abishek Gopalan, Soshant Bali, Mikel Jimenez Fernandez, and Grace Smith. 2021. Network hose: Managing uncertain network demand with model simplicity. https://engineering.fb.com/2021/06/15/data-infrastructure/network-hose/.Google Scholar
- Reid Andersen, Fan Chung, Arunabha Sen, and Guoliang Xue. 2004. On disjoint path pairs with wavelength continuity constraint in WDM networks. In IEEE INFOCOM 2004, Vol. 1. IEEE.Google ScholarCross Ref
- Srivatsan Balasubramanian, Vinayak Dangui, John P Eason, and Satyajeet Singh Ahuja. 2023. Targeted Defragmentation of a Production Optical Network. In 2023 International Conference on Optical Network Design and Modeling (ONDM). IEEE, 1--3.Google Scholar
- Hitesh Ballani, Paolo Costa, Thomas Karagiannis, and Ant Rowstron. 2011. Towards predictable datacenter networks. In Proceedings of the ACM SIGCOMM 2011 Conference. 242--253.Google ScholarDigital Library
- Walid Ben-Ameur and José Neto. 2007. Acceleration of cutting-plane and column generation algorithms: Applications to network design. Networks: An International Journal 49, 1 (2007), 3--17.Google Scholar
- Timo Berthold. 2006. Primal heuristics for mixed integer programs. Ph.D. Dissertation. Zuse Institute Berlin (ZIB).Google Scholar
- Haesun Byun and Meejeong Lee. 2007. Extensions to P2MP RSVP-TE for VPN-specific state provisioning with fair resource sharing. Computer communications 30, 18 (2007), 3736--3745.Google Scholar
- Richard Li-Yang Chen, Amy Cohn, and Ali Pinar. 2011. An implicit optimization approach for survivable network design. In 2011 IEEE network science workshop. IEEE, 180--187.Google Scholar
- Qi Deng, Galen Sasaki, and Ching-Fong Su. 2002. Survivable IP over WDM: an efficient mathematical programming problem formulation. In PROCEEDINGS OF THE ANNUAL ALLERTON CONFERENCE ON COMMUNICATION CONTROL AND COMPUTING, Vol. 40. 518--527.Google Scholar
- Nick G Duffield, Pawan Goyal, Albert Greenberg, Partho Mishra, Kadangode K Ramakrishnan, and Jacobus E van der Merive. 1999. A flexible model for resource management in virtual private networks. In Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication. 95--108.Google ScholarDigital Library
- Friedrich Eisenbrand and Edda Happ. 2006. Provisioning a virtual private network under the presence of non-communicating groups. In Italian Conference on Algorithms and Complexity. Springer, 105--114.Google ScholarDigital Library
- J Andrew Fingerhut, Subhash Suri, and Jonathan S Turner. 1997. Designing least-cost nonblocking broadband networks. Journal of Algorithms 24, 2 (1997), 287--309.Google ScholarDigital Library
- Albert Greenberg, James R. Hamilton, Navendu Jain, Srikanth Kandula, Changhoon Kim, Parantap Lahiri, David A. Maltz, Parveen Patel, and Sudipta Sengupta. 2009. VL2: A scalable and flexible data center network. In In Proceedings of the ACM SIGCOMM 2009 conference on Data communication. 51--62.Google ScholarDigital Library
- John N Hooker and Greger Ottosson. 2003. Logic-based Benders decomposition. Mathematical Programming 96, 1 (2003), 33--60.Google ScholarCross Ref
- Simon Kassing, Asaf Valadarsky, Gal Shahaf, Michael Schapira, and Ankit Singla. 2017. Beyond fat-trees without antennae, mirrors, and disco-balls. In Proceedings of the Conference of the ACM Special Interest Group on Data Communication. 281--294.Google ScholarDigital Library
- Kayi Lee, Eytan Modiano, and Hyang-Won Lee. 2010. Cross-layer survivability in WDM-based networks. IEEE/ACM Transactions on Networking 19, 4 (2010), 1000--1013.Google ScholarDigital Library
- Dion Leung and Wayne D Grover. 2005. Capacity planning of survivable mesh-based transport networks under demand uncertainty. Photonic Network Communications 10 (2005), 123--140.Google ScholarCross Ref
- Sara Mattia. 2013. The robust network loading problem with dynamic routing. Computational Optimization and Applications 54, 3 (2013), 619--643.Google ScholarDigital Library
- Biswanath Mukherjee, Dhritiman Banerjee, Sav Ramamurthy, and Amarnath Mukherjee. 1996. Some principles for designing a wide-area WDM optical network. IEEE/ACM transactions on networking 4, 5 (1996), 684--696.Google ScholarDigital Library
- Ragheb Rahmaniani, Teodor Gabriel Crainic, Michel Gendreau, and Walter Rei. 2017. The Benders decomposition algorithm: A literature review. European Journal of Operational Research 259, 3 (2017), 801--817.Google ScholarCross Ref
- Rajiv Ramaswami and Kumar N Sivarajan. 1996. Design of logical topologies for wavelength-routed optical networks. IEEE Journal on Selected areas in communications 14, 5 (1996), 840--851.Google ScholarDigital Library
- Henrique Rodrigues, Jose Renato Santos, Yoshio Turner, Paolo Soares, and Dorgival Guedes. 2011. Gatekeeper: Supporting bandwidth guarantees for multi-tenant datacenter networks. In 3rd Workshop on I/O Virtualization (WIOV 11).Google Scholar
- Mike Shand and Stewart Bryant. 2010. IP fast reroute framework. Technical Report.Google Scholar
- Z Caner Taskın. 2010. Benders decomposition. Wiley Encyclopedia of Operations Research and Management Science. John Wiley & Sons, Malden (MA) (2010).Google Scholar
Index Terms
- Hose-based cross-layer backbone network design with Benders decomposition
Recommendations
Capacity-efficient and uncertainty-resilient backbone network planning with hose
SIGCOMM '21: Proceedings of the 2021 ACM SIGCOMM 2021 ConferenceThis paper presents Facebook's design and operational experience of a Hose-based backbone network planning system. This initial adoption of the Hose model in network planning is driven by the capacity and demand uncertainty pressure of backbone ...
Exact algorithms based on Benders decomposition for multicommodity uncapacitated fixed-charge network design
Highlights- We combine mixed integer programming tools such as Benders decomposition, branch-and-cut, lift-and-project cuts, matheuristics, cut-and-solve, and the ...
AbstractThis paper presents two exact algorithms based on Benders decomposition for solving the multicommodity uncapacitated fixed-charge network design problem. The first is a branch-and-cut algorithm based on a Benders reformulation enhanced ...
Near-optimal tree-based access network design
Among various access network topologies, the tree topology is the most popular due to its simplicity and relatively low cost. A salient example is the CATV network. In this paper, we consider the tree-based access network design problem where the ...
Comments