Poster - Dissecting Distributed Database Systems For Telecom Applications
May 15, 2016
·
0 min read
Abstract
This poster outlines a collaborative research project by Ericsson and ELTE University, focused on analyzing MySQL NDB cluster communication. In this project, we developed methods for granular analysis of cluster response times, involving an in-depth examination of the communication protocol. Despite the absence of comprehensive protocol documentation, insights were drawn from MySQL Cluster source code and real CUDB system packet traces. Examination of the response times of some key queries opens the possibility for dynamic load balancing.
Date
May 15, 2016 2:00 PM — 3:00 PM
Event
Ericsson University Day
Location
Eotvos Lorand University
1-3 Egyetem tér, Budapest, 1053
Authors
Quantitative Researcher
Quantitative researcher with a PhD in Mathematics, combining expertise in stochastic modeling, statistical inference, and machine learning with industrial experience in developing predictive systems.
Experienced in large-scale time-series analysis, uncertainty quantification, and rigorous model validation, with strong Python/C++ skills and a proven track record of translating research into production software.
Currently developing intraday electricity-price models and stochastic-control methods for optimal liquidation at Aarhus University, in collaboration with an energy trading firm.