License
When quoting this document, please refer to the following
DOI: 10.4230/LIPIcs.APPROX-RANDOM.2015.381
URN: urn:nbn:de:0030-drops-53133
URL: http://drops.dagstuhl.de/opus/volltexte/2015/5313/
Go to the corresponding LIPIcs Volume Portal


Iglesias, Jennifer ; Rajaraman, Rajmohan ; Ravi, R. ; Sundaram, Ravi

Designing Overlapping Networks for Publish-Subscribe Systems

pdf-format:
23.pdf (0.5 MB)


Abstract

From the publish-subscribe systems of the early days of the Internet to the recent emergence of Web 3.0 and IoT (Internet of Things), new problems arise in the design of networks centered at producers and consumers of constantly evolving information. In a typical problem, each terminal is a source or sink of information and builds a physical network in the form of a tree or an overlay network in the form of a star rooted at itself. Every pair of pub-sub terminals that need to be coordinated (e.g. the source and sink of an important piece of control information) define an edge in a bipartite demand graph; the solution must ensure that the corresponding networks rooted at the endpoints of each demand edge overlap at some node. This simple overlap constraint, and the requirement that each network is a tree or a star, leads to a variety of new questions on the design of overlapping networks. In this paper, for the general demand case of the problem, we show that a natural LP formulation has a non-constant integrality gap; on the positive side, we present a logarithmic approximation for the general demand case. When the demand graph is complete, however, we design approximation algorithms with small constant performance ratios, irrespective of whether the pub networks and sub networks are required to be trees or stars.

BibTeX - Entry

@InProceedings{iglesias_et_al:LIPIcs:2015:5313,
  author =	{Jennifer Iglesias and Rajmohan Rajaraman and R. Ravi and Ravi Sundaram},
  title =	{{Designing Overlapping Networks for Publish-Subscribe Systems}},
  booktitle =	{Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques (APPROX/RANDOM 2015)},
  pages =	{381--395},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-939897-89-7},
  ISSN =	{1868-8969},
  year =	{2015},
  volume =	{40},
  editor =	{Naveen Garg and Klaus Jansen and Anup Rao and Jos{\'e} D. P. Rolim},
  publisher =	{Schloss Dagstuhl--Leibniz-Zentrum fuer Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{http://drops.dagstuhl.de/opus/volltexte/2015/5313},
  URN =		{urn:nbn:de:0030-drops-53133},
  doi =		{10.4230/LIPIcs.APPROX-RANDOM.2015.381},
  annote =	{Keywords: Approximation Algorithms, Steiner Trees, Publish-Subscribe Systems, Integrality Gap, VPN.}
}

Keywords: Approximation Algorithms, Steiner Trees, Publish-Subscribe Systems, Integrality Gap, VPN.
Seminar: Approximation, Randomization, and Combinatorial Optimization. Algorithms and Techniques (APPROX/RANDOM 2015)
Issue Date: 2015
Date of publication: 28.07.2015


DROPS-Home | Fulltext Search | Imprint Published by LZI