Bio-Inspired Techniques for Resources State Prediction in Large Scale Distributed Systems

Abstract

The state prediction of resources in large scale distributed systems represents an important aspect for resources allocations, systems evaluation, and autonomic control. The paper presents advanced techniques for resources state prediction in Large Scale Distributed Systems, which include techniques based on bio-inspired algorithms like neural network improved with genetic algorithms. The approach adopted in this paper consists of a new fitness function, having prediction error minimization as the main scope. The proposed prediction techniques are based on monitoring data, aggregated in a history database. The experimental scenarios consider the ALICE experiment, active at the CERN institute. Compared with classical predicted algorithms based on average or random methods, the authors obtain an improved prediction error of 73%. This improvement is important for functionalities and performance of resource management systems in large scale distributed systems in the case of remote control ore advance reservation and allocation.

A methodology for assessing the predictable behaviour of mobile users in wireless networks

“…As physical location has gradually become a stereotype in context‐aware computing, a need for more complex context is already expressed by various studies . In , Hui and Crowcroft study the impact of predictable human interactions on forwarding in Pocket Switched Networks.…”

Simulation of SLA-based VM-scaling algorithms for cloud-distributed applications

“…Visan et al [11] describe a bio-inspired prediction algorithm based on a Cascade-Correlation neural network, which uses a genetic algorithm for initialising the network's weights. The authors use their algorithm for performing both one-step and multi-step predictions of a large-scale distributed experiment, with good results.…”

“…The predictions can then be used for optimising processes with longer time horizons [11], such as the allocation of physical computing resources [14]. For the work presented in Section 6 we will use a combination of linear models and error estimation, as the time horizon of the prediction is usually short, in the range of tens of seconds.…”

Service level agreements-driven management of distributed applications in cloud computing environments

“…Visan et al [148] describe a bio-inspired prediction algorithm based on a Cascade-Correlation neural network, which uses a genetic algorithm for initialising the network's weights. The authors use their algorithm for performing both one-step and multi-step predictions of a large-scale distributed experiment, with good results.…”

“…The lower image shows a bad prediction caused by the incorrect detection of the trend of the time series, caused by the shorter value of the period -the signal being The non-linear prediction models are suited for forecasting time series with seasonal variations, such as hourly, daily or monthly. The predictions can then be used for optimising processes with longer time horizons [148], such as the allocation of physical computing resources [19]. For the work presented in Section 4.4 we will use a combination of linear models and error estimation, as the time horizon of the prediction is usually short, in the range of tens of seconds.…”