Demand-aware network designs of bounded degree
From MaRDI portal
Publication:2189175
DOI10.1007/S00446-019-00351-5zbMATH Open1445.68027arXiv1705.06024OpenAlexW2614467483MaRDI QIDQ2189175FDOQ2189175
Authors: Chen Avin, Kaushik Mondal, Stefan Schmid 
Publication date: 15 June 2020
Published in: Distributed Computing (Search for Journal in Brave)
Abstract: Traditionally, networks such as datacenter interconnects are designed to optimize worst-case performance under arbitrary traffic patterns. Such network designs can however be far from optimal when considering the actual workloads and traffic patterns which they serve. This insight led to the development of demand-aware datacenter interconnects which can be reconfigured depending on the workload. Motivated by these trends, this paper initiates the algorithmic study of demand-aware networks (DANs) designs, and in particular the design of bounded-degree networks. The inputs to the network design problem are a discrete communication request distribution, D, defined over communicating pairs from the node set V , and a bound, d, on the maximum degree. In turn, our objective is to design an (undirected) demand-aware network N = (V,E) of bounded-degree d, which provides short routing paths between frequently communicating nodes distributed across N. In particular, the designed network should minimize the expected path length on N (with respect to D), which is a basic measure of the efficiency of the network. We show that this fundamental network design problem exhibits interesting connections to several classic combinatorial problems and to information theory. We derive a general lower bound based on the entropy of the communication pattern D, and present asymptotically optimal network-aware design algorithms for important distribution families, such as sparse distributions and distributions of locally bounded doubling dimensions.
Full work available at URL: https://arxiv.org/abs/1705.06024
Recommendations
Graph theory (including graph drawing) in computer science (68R10) Distributed algorithms (68W15) Network design and communication in computer systems (68M10)
Cites Work
- Introduction to algorithms.
- Fast Construction of Nets in Low-Dimensional Metrics and Their Applications
- Optimal Assignments of Numbers to Vertices
- Low distortion spanners
- Nearly-linear time algorithms for graph partitioning, graph sparsification, and solving linear systems
- Self-adjusting binary search trees
- An Optimal Synchronizer for the Hypercube
- A history of graph entropy measures
- Graph spanners
- Additive spanners and \(({\alpha}, {\beta})\)-spanners
- Sparse Sourcewise and Pairwise Distance Preservers
- Fast Estimation of Diameter and Shortest Paths (Without Matrix Multiplication)
- The complexity of the network design problem
- Spanners with Slack
- An improved approximation ratio for the minimum linear arrangement problem
- Nearly optimal binary search trees
- Minimizing Average Shortest Path Distances via Shortcut Edge Addition
- \(\ell ^2_2\) spreading metrics for vertex ordering problems
- The geometry of binary search trees
- Chordal completions of planar graphs
- Dynamic Optimality—Almost
- $(1 + \epsilon,\beta)$-Spanner Constructions for General Graphs
- O(log log n)-competitive dynamic binary search trees
- A Doubling Dimension Threshold Θ(loglogn) for Augmented Graph Navigability
- Spanners and emulators with sublinear distance errors
- rDAN: toward robust demand-aware network designs
- Small hop-diameter sparse spanners for doubling metrics
- On hierarchical routing in doubling metrics
- Self-adjusting grid networks to minimize expected path length
- Embedding Graphs with Bounded Treewidth into Their Optimal Hypercubes
Cited In (4)
This page was built for publication: Demand-aware network designs of bounded degree
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2189175)
