## Provisioning a Virtual Private Network: A network design problem for multicommodity flow (2001)

Venue: | In Proceedings of the 33rd Annual ACM Symposium on Theory of Computing |

Citations: | 83 - 13 self |

### BibTeX

@INPROCEEDINGS{Gupta01provisioninga,

author = {Anupam Gupta and Jon Kleinberg and Amit Kumar and Rajeev Rastogi and Bulent Yener},

title = {Provisioning a Virtual Private Network: A network design problem for multicommodity flow},

booktitle = {In Proceedings of the 33rd Annual ACM Symposium on Theory of Computing},

year = {2001},

pages = {389--398}

}

### Years of Citing Articles

### OpenURL

### Abstract

Consider a setting in which a group of nodes, situated in a large underlying network, wishes to reserve bandwidth on which to support communication. Virtual private networks (VPNs) are services that support such a construct; rather than building a new physical network on the group of nodes that must be connected, bandwidth in the underlying network is reserved for communication within the group, forming a virtual “sub-network.” Provisioning a virtual private network over a set of terminals gives rise to the following general network design problem. We have bounds on the cumulative amount of traffic each terminal can send and receive; we must choose a path for each pair of terminals, and a bandwidth allocation for each edge of the network, so that any traffic matrix consistent with the given upper bounds can be feasibly routed. Thus, we are seeking to design a network that can support a continuum of possible traffic scenarios. We provide optimal and approximate algorithms for several variants of this problem, depending on whether the traffic matrix is required to be symmetric, and on whether the designed network is required to be a tree (a natural constraint in a number of basic applications). We also establish a relation between this collection of network design problems and a variant of the facility location problem introduced by Karger and Minkoff; we extend their results by providing a stronger approximation algorithm for this latter problem. 1

### Citations

11205 |
Computers and Intractability: a Guide to the Theory of NP-completeness
- Garey, Johnson
- 1979
(Show Context)
Citation Context ...out to be NP-hard. Theorem 2.4 The problem� is strongly NP-hard and max-SNP hard. Proof: We give an approximation-preserving reduction from the minimum Steiner tree problem, which is strongly NP-hard =-=[9]-=-. Given a Steiner tree instance on the graph�� ��with edge weights� and required vertices , we create an instance of� , where the edge weights are�, is the set of terminals, each with thresholds����� ... |

1164 | Geometric Algorithms and Combinatorial Optimization - Lovász, Grötschel, et al. - 1993 |

522 | Matching Theory - Lovász, Plummer - 1986 |

322 | Approximation algorithms for metric facility location and kmedian problems using the primal-dual scheme and lagrangian relaxation - JAIN, VAZIRANI |

267 | Approximation algorithms for facility location problems
- Shmoys, Tardos, et al.
- 1997
(Show Context)
Citation Context ...a constant factor. We provide a significantly improved constant-factor approximation algorithm for the Connected Facility Location problem, adapting a rounding technique of Shmoys, Tardos, and Aardal =-=[22]-=-. Combined with our underlying reduction, this yields an approximation algorithm for� . Although we arrived at the Connected Facility Location problem as a step in approximating network design, we fee... |

206 | Improved combinatorial algorithms for facility location problems - Charikar, Guha |

202 | A factor 2 approximation algorithm for the generalized Steiner network problem
- Jain
- 1998
(Show Context)
Citation Context ...ed to algorithmic work on two other network design models, namely survivable network design and buy-at-bulk network design, but with some crucial differences. In survivable network design (see, e.g., =-=[10, 13, 23, 24]-=-), we are given a graph�with integer-valued flow requirements���between certain pairs of nodes�and�. We must choose a minimum-cost subgraph of�so that for each pair���, there are at least���edge-disjo... |

198 | Improved steiner tree approximation in graphs
- Robins, Zelikovsky
(Show Context)
Citation Context ...threshold �, the arguments of Theorem 2.4 show that the optimal tree solution corresponds exactly to the optimal Steiner tree connecting the terminal set, and hence can be approximated to within 1.55 =-=[19]-=-. � � factor 3 Unsplittable and Fractional Graph Solutions 3.1 The Symmetric Case In this section, we show that a solution to is a constant factor approximation to �. In fact, we show something strong... |

100 | Buy-at-bulk network design
- Awerbuch, Azar
- 1997
(Show Context)
Citation Context ...mmodity network design problem; the subgraph constructed needs to satisfy the demand for any single pair of nodes, but not for multiple pairs simultaneously. In buy-at-bulk network design (see, e.g., =-=[2, 4, 20]-=-), there is a specified demand���between every pair of nodes in a graph�, and we must buy capacity on edges as cheaply as possible so as to support a (fractional or integral) multicommodity flow satis... |

97 |
MPLS: Technology and Applications
- Davie, Rekhter
- 2000
(Show Context)
Citation Context ..., since fewer labels have 2P 12 3 P 12 3 1 P 23 1 P 23 2 2 P 13 P 13 Figure 1: Two possible sets of paths for a 3-terminal symmetric instance. to be used, and label stacks on packets are not as deep =-=[6]-=-. Furthermore, trees simplify restoration of paths in case of link failures, since all paths traversing a failed link can be restored as a single group, instead of each path being restored separately.... |

84 | A primal-dual approximation algorithm for generalized Steiner network problems
- Williamson, Goemans, et al.
- 1995
(Show Context)
Citation Context ...ed to algorithmic work on two other network design models, namely survivable network design and buy-at-bulk network design, but with some crucial differences. In survivable network design (see, e.g., =-=[10, 13, 23, 24]-=-), we are given a graph�with integer-valued flow requirements���between certain pairs of nodes�and�. We must choose a minimum-cost subgraph of�so that for each pair���, there are at least���edge-disjo... |

79 | Improved approximation algorithms for network design problems
- Goemans, Goldberg, et al.
- 1994
(Show Context)
Citation Context ...ed to algorithmic work on two other network design models, namely survivable network design and buy-at-bulk network design, but with some crucial differences. In survivable network design (see, e.g., =-=[10, 13, 23, 24]-=-), we are given a graph�with integer-valued flow requirements���between certain pairs of nodes�and�. We must choose a minimum-cost subgraph of�so that for each pair���, there are at least���edge-disjo... |

77 | Building Steiner trees with incomplete global knowledge
- Karger, Minkoff
- 2000
(Show Context)
Citation Context ...we must also choose a Steiner tree which connects all the open facilities, and we pay an additional cost proportional to the total edge length of . 3This problem was introduced by Karger and Minkoff =-=[15]-=-, who developed an algorithm approximating the optimum to within a constant factor. We provide a significantly improved constant-factor approximation algorithm for the Connected Facility Location prob... |

71 | Approximation algorithms for geometric median problems
- LIN, VITTER
- 1992
(Show Context)
Citation Context ... we can get an improved approximation guarantee of 10. Proof: The basic idea of the rounding algorithm is similar to the one given by Shmoys et al. [22]. The algorithm first performs a filtering step =-=[16, 17]-=- followed by clustering, after which a facility is chosen from each cluster. By the properties of the filtering, the cost of opening this facility and connecting the demand points to it is not much mo... |

68 | Hierarchical placement and network design problems
- Guha, Meyerson, et al.
- 2000
(Show Context)
Citation Context ...nt-factor approximation algorithm. Their approach works in two stages. The first stage decides which facilities to open by “clustering” demands as much as possible (see also the paper by Guha et. al. =-=[12]-=-). In the second stage, they prove that it does not cost much to connect these facilities by a minimum Steiner tree. Our algorithm formulates a single linear programming relaxation for the combined pr... |

62 |
der Merwe. A flexible model for resource management in virtual private networks
- Duffield, Goyal, et al.
- 1999
(Show Context)
Citation Context ... than building a new physical network on the group of nodes that must be connected, bandwidth in the underlying network is reserved for communication within the group, forming a virtual “sub-network” =-=[7, 8, 21]-=-. A preliminary version of this paper appeared in the Proceedings of the 33nd Annual ACM Symposium on Theory of Computing, 2001[?]. Bell Labs, 700 Mountain Avenue, Murray Hill NJ 07974. Email: anupamg... |

45 | The access network design problem
- MatthewAndrews, Zhang
- 1998
(Show Context)
Citation Context ...mmodity network design problem; the subgraph constructed needs to satisfy the demand for any single pair of nodes, but not for multiple pairs simultaneously. In buy-at-bulk network design (see, e.g., =-=[2, 4, 20]-=-), there is a specified demand���between every pair of nodes in a graph�, and we must buy capacity on edges as cheaply as possible so as to support a (fractional or integral) multicommodity flow satis... |

34 |
Buy-at-bulk network design: Approximating the single-sink edge installation problem
- Salman, Cheriyan, et al.
- 1997
(Show Context)
Citation Context ...mmodity network design problem; the subgraph constructed needs to satisfy the demand for any single pair of nodes, but not for multiple pairs simultaneously. In buy-at-bulk network design (see, e.g., =-=[2, 4, 20]-=-), there is a specified demand���between every pair of nodes in a graph�, and we must buy capacity on edges as cheaply as possible so as to support a (fractional or integral) multicommodity flow satis... |

32 |
approximations with minimum packing constraint violation (extended abstract
- Lin, Vitter
- 1992
(Show Context)
Citation Context ... we can get an improved approximation guarantee of 10. Proof: The basic idea of the rounding algorithm is similar to the one given by Shmoys et al. [22]. The algorithm first performs a filtering step =-=[16, 17]-=- followed by clustering, after which a facility is chosen from each cluster. By the properties of the filtering, the cost of opening this facility and connecting the demand points to it is not much mo... |

30 | A 3-approximation algorithm for the k-level uncapacitated facility location problem - Aardal, Chudak, et al. - 1999 |

18 |
ATM virtual private networks
- Fotedar, Gerla, et al.
- 1995
(Show Context)
Citation Context ... than building a new physical network on the group of nodes that must be connected, bandwidth in the underlying network is reserved for communication within the group, forming a virtual “sub-network” =-=[7, 8, 21]-=-. A preliminary version of this paper appeared in the Proceedings of the 33nd Annual ACM Symposium on Theory of Computing, 2001[?]. Bell Labs, 700 Mountain Avenue, Murray Hill NJ 07974. Email: anupamg... |

17 | Approximating the tree and tour covers of a graph - Arkin, Halldórsson, et al. - 1993 |

11 |
Approximation Algorithms for a Class of Graph Problems
- Williamson
- 1993
(Show Context)
Citation Context |

9 | Charikar and Sudipto Guha. Improved combinatorial algorithms for the facility location and k-median problems - Moses - 1999 |

6 |
Virtual Private Networks. O’Reilly
- Scott, Wolfe, et al.
- 1998
(Show Context)
Citation Context ... than building a new physical network on the group of nodes that must be connected, bandwidth in the underlying network is reserved for communication within the group, forming a virtual “sub-network” =-=[7, 8, 21]-=-. A preliminary version of this paper appeared in the Proceedings of the 33nd Annual ACM Symposium on Theory of Computing, 2001[?]. Bell Labs, 700 Mountain Avenue, Murray Hill NJ 07974. Email: anupamg... |

3 | Awerbuch and Yossi Azar. Buy-at-bulk network design - Baruch - 1997 |

2 | approximation algorithms for network design problems - Improved - 1994 |

1 |
A 3-approximation algorithm for the�-level uncapacitated facility location problem
- Aardal, Chudak, et al.
- 1999
(Show Context)
Citation Context ...ding to a significantly improved constant factor. The connected facility location problem is also related to other facility location variants, including multi-level and loadbalanced facility location =-=[1, 12]-=-, as well as to the problem of finding connected vertex covers [3]. The main network design problems considered here are related to algorithmic work on two other network design models, namely survivab... |