QoS modeling and analysis of component‐based software systems: a stochastic approach

Abstract

SUMMARYThere is a growing demand for using commercial-off-the-shelf (COTS) software components to facilitate the development of software systems. Among many research topics for component-based software, qualityof-service (QoS) evaluation is yet to be given the importance it deserves. In this paper, we propose a novel analytical model to evaluate the QoS of component-based software systems. We use the component execution graph (CEG) graph model to model the architecture at the process level and the interdependence among components. The CEG graph can explicitly capture sequential, parallel, selective and iterative compositions of components. For QoS estimation, each component in the CEG model is associated with execution rate, failure rate and cost per unit time. Three metrics of the QoS are considered and analytically calculated, namely make-span, reliability and cost. Through a case study, we show that our model is capable of modeling real-world COTS software systems effectively. Also, Monte-Carlo simulation in the case study indicates that analytical results are consistent with simulation and all are covered by 95% confidence intervals. We also present a sensitivity analysis technique to identify QoS bottlenecks. This paper concludes with a comparison with related work.

A Petri-Net-Based Approach to Reliability Determination of Ontology-Based Service Compositions

“…Considering that the theoretical distribution of NRE is difficult to evaluate (several studies suggest that the probability of skipping repetition when every iteration ends can be approximately stable, and the number of repetitive cycles therefore follows the geometrical distribution and employs the expectation of geometrical distribution in evaluating the QoS of repetitive constructs, e.g., [14] and [32]- [34]), we can approximately calculate Prob{NRE = i} as the occurrence rate of NRE equaling i based on the experimental test or historical runtime logs. That is,…”

Reliability Prediction of Ontology-Based Service Compositions Using Petri Net and Time Series Models

“…Letting NRE denote the number of repetitive cycles to achieve completion (by marking the completed place in Figure 8 ), we have that the reliability for 〈 repeat - while 〉, RRW , is calculated as the total probability of where P { NRE = i } denotes the probability that NRE equals i . Since the theoretical distribution of NRE is difficult to evaluate (some studies suggest that the probability of skipping repetition when every iteration ends can be approximately seen stable and the number of repetitive cycles therefore follows the geometrical distribution and employ the expectation of geometrical distribution in evaluating the QoS of repetitive constructs, e.g., [ 4 , 16 , 17 ]), we can approximately calculate P { NRE = i } as the occurrence rate of NRE equalling i based on the experimental test or historical logs. That is where frq ( NRE equalling i ) denotes the frequency of the event of NRE equalling i and NT denotes the number of experimental trials.…”