Decentralizing Network Inference Problems With Multiple-Description Fusion Estimation (MDFE)
Abstract
Network inference (or tomography) problems, such as traffic matrix estimation or completion and link loss inference, have been studied rigorously in different networking applications. These problems are often posed as under-determined linear inverse (UDLI) problems and solved in a centralized manner, where all the measurements are collected at a central node, which then applies a variety of inference techniques to estimate the attributes of interest. This paper proposes a novel framework for decentralizing these large-scale under-determined network inference problems by intelligently partitioning it into smaller subproblems and solving them independently and in parallel. The resulting estimates, referred to as multiple descriptions, can then be fused together to compute the global estimate. We apply this Multiple Description and Fusion Estimation (MDFE) framework to three classical problems: traffic matrix estimation, traffic matrix completion, and loss inference. Using real topologies and traces, we demonstrate how MDFE can speed up computation while maintaining (even improving) the estimation accuracy and how it enhances robustness against noise and failures. In conclusion, we also show that our MDFE framework is compatible with a variety of existing inference techniques used to solve the UDLI problems.
- Authors:
-
- Univ. of California, Davis, CA (United States)
- HP-Labs, Palo Alto, CA (United States)
- Publication Date:
- Research Org.:
- Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC)
- OSTI Identifier:
- 1525161
- Grant/Contract Number:
- AC02-05CH11231
- Resource Type:
- Accepted Manuscript
- Journal Name:
- IEEE/ACM Transactions on Networking
- Additional Journal Information:
- Journal Volume: 24; Journal Issue: 4; Journal ID: ISSN 1063-6692
- Publisher:
- IEEE - ACM
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 97 MATHEMATICS AND COMPUTING; distributed estimation; multiple description fusion; network inference; traffic matrix estimation; under-determined inverse problem
Citation Formats
Malboubi, Mehdi, Vu, Cuong, Chuah, Chen -Nee, and Sharma, Puneet. Decentralizing Network Inference Problems With Multiple-Description Fusion Estimation (MDFE). United States: N. p., 2015.
Web. doi:10.1109/TNET.2015.2475362.
Malboubi, Mehdi, Vu, Cuong, Chuah, Chen -Nee, & Sharma, Puneet. Decentralizing Network Inference Problems With Multiple-Description Fusion Estimation (MDFE). United States. https://doi.org/10.1109/TNET.2015.2475362
Malboubi, Mehdi, Vu, Cuong, Chuah, Chen -Nee, and Sharma, Puneet. Thu .
"Decentralizing Network Inference Problems With Multiple-Description Fusion Estimation (MDFE)". United States. https://doi.org/10.1109/TNET.2015.2475362. https://www.osti.gov/servlets/purl/1525161.
@article{osti_1525161,
title = {Decentralizing Network Inference Problems With Multiple-Description Fusion Estimation (MDFE)},
author = {Malboubi, Mehdi and Vu, Cuong and Chuah, Chen -Nee and Sharma, Puneet},
abstractNote = {Network inference (or tomography) problems, such as traffic matrix estimation or completion and link loss inference, have been studied rigorously in different networking applications. These problems are often posed as under-determined linear inverse (UDLI) problems and solved in a centralized manner, where all the measurements are collected at a central node, which then applies a variety of inference techniques to estimate the attributes of interest. This paper proposes a novel framework for decentralizing these large-scale under-determined network inference problems by intelligently partitioning it into smaller subproblems and solving them independently and in parallel. The resulting estimates, referred to as multiple descriptions, can then be fused together to compute the global estimate. We apply this Multiple Description and Fusion Estimation (MDFE) framework to three classical problems: traffic matrix estimation, traffic matrix completion, and loss inference. Using real topologies and traces, we demonstrate how MDFE can speed up computation while maintaining (even improving) the estimation accuracy and how it enhances robustness against noise and failures. In conclusion, we also show that our MDFE framework is compatible with a variety of existing inference techniques used to solve the UDLI problems.},
doi = {10.1109/TNET.2015.2475362},
journal = {IEEE/ACM Transactions on Networking},
number = 4,
volume = 24,
place = {United States},
year = {Thu Sep 17 00:00:00 EDT 2015},
month = {Thu Sep 17 00:00:00 EDT 2015}
}
Web of Science
Figures / Tables:
Figures / Tables found in this record: