Prediction of Distributed Systems State Based on Monitoring Data

A. Draghici,Alexandru Costan,V. Cristea
DOI: https://doi.org/10.1109/ISPDC.2010.28
2010-07-07
Abstract:Autonomic behavior has emerged as a solution for the issues related to performance improvement and resource-usage optimization in large scale distributed systems. This solution relies on monitoring services to keep track of the states of the managed systems. However, most of the monitoring services are designed to provide general resource information and do not consider specific information for higher-level services, lacking important control capabilities. In this context, a dynamic adaptation layer is required. Based on the collected monitoring information in conjunction with some planning and prediction algorithms, it should be able to reactive and proactive deal with detected or predicted conditions. This paper presents a prediction architecture developed within the MonALISA monitoring framework, providing methods for estimating future values for different parameters on various periods of time. The predictions are used to enhance the self-adaptive behavior of several data intensive applications. Our research was focused on machine learning algorithms correlated with statistical techniques for data mining purposes in order to perform n-step-ahead time series predictions and to evaluate their performances dynamical.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to calculate the light hadron spectrum using the Wilson action within the framework of quenched quantum chromodynamics (QCD) and determine the masses of these particles as accurately as possible. Specifically, the authors focus on the following aspects: 1. **Accurate measurement of light hadron masses**: Researchers hope to obtain more accurate masses of light hadrons (such as mesons and baryons) through simulations under different coupling constants \(\beta\) and quark mass ranges. 2. **Quark mass dependence**: Explore how hadron masses change with quark masses, especially the behavior when the quark mass is close to the physical value. For example, they find that the nucleon mass exhibits non - linear behavior as the quark mass decreases, which is different from previous expectations. 3. **Continuum limit extrapolation**: By performing simulations at different lattice spacings \(a\), researchers attempt to extrapolate the results to the continuum limit (i.e., the lattice spacing approaches zero) to reduce the influence of discretization errors on the results. 4. **Determination of quark masses**: Determine the masses of light quarks by two methods, the axial Ward identity and perturbation theory, and compare the results of the two to ensure their consistency. 5. **Calculation of decay constants**: Calculate the decay constants \(f_V\) and \(f_P\) for vector mesons and pseudoscalar mesons, and evaluate the effects of different renormalization schemes. 6. **Influence of strange quark effects**: Study the influence of strange quarks on the masses of mesons and baryons, and find that the masses of strange hadrons are systematically lower than the experimental values, indicating some problems in the confinement approximation. In summary, this paper aims to gain an in - depth understanding of the behavior of the light hadron spectrum in confined QCD through high - precision numerical simulations and reveal possible theoretical deviations or deficiencies. This will help to improve future QCD calculation models and make them closer to experimental observations.