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.
Location
Eotvos Lorand University
1-3 Egyetem tér, Budapest, 1053
Authors
Quantitative Researcher
Quantitative Researcher with a PhD in Mathematics, specializing in stochastic modeling, machine learning, and predictive systems for financial markets.
Experienced in probabilistic modeling, Monte Carlo simulation, uncertainty quantification, and statistical validation for data-driven decision-making.
Currently developing intraday energy-market price prediction models and optimal liquidation strategies using machine learning, functional data analysis, and stochastic differential equations.
Interested in market prediction problems where model quality is directly reflected in trading performance and PnL.