Workload Characterization for Software Aging Analysis

Abstract

The phenomenon of software aging is increasingly recognized as a relevant problem of long-running systems. Numerous experiments have been carried out in the last decade to empirically analyze software aging. Such experiments, besides highlighting the relevance of the phenomenon, have shown that aging is tightly related to the applied workload. However, due to the differences among the experimented applications and among the experimental conditions, results of past studies are not comparable to each other. This prevent from drawing general conclusions (e.g., about the aging-workload relationship), and from comparing systems from the aging perspective. In this paper, we propose a procedure to carry out aging experiments in different applications for: i) assessing aging trend of the individual systems, as well as assessing differences among them (i.e., obtaining comparable results), ii) inferring workload-aging relationships from experiments performed on different applications, by highlighting the most relevant workload parameters. The procedure is applied, through a set of long-running experiments, to three real-scale software applications, namely Apache Web Server, James Mail Server, and CARDAMOM, a middleware for the development of air traffic control (ATC) systems

On the Aging Effects Due to Concurrency Bugs: A Case Study on MySQL

“…Since aging has been shown to be clearly correlated with workload variation, several authors also accounted for its impact, highlighting its dependency on aging effects and thus on the failure rate. Examples are in [2], [14], [22], where the estimate of the time-to-aging-failure at a given time varies in function of the workload actually experienced by the system. Attempts to study aging from a different perspective are in the works by Matias et.…”

“…To take the workload influence into account, different parameters have been considered to analyze MySQL aging. Following the procedure proposed in our previous study [22], we characterize MySQL workload with the following parameters:…”

“…Moreover, to avoid failures due to excessive stressful workload (which is not related to aging failures), we also need to determine what is the maximum capacity for the system under test. To these aims, capacity tests are performed, according to the procedure described in our previous work [22], in order to define the limits of MySQL server deployed in our testbed. To evaluate the actual limit (i.e., maximum number of requests correctly served in a time unit), we need to take into account the levels of other workload parameters (such as request types and exchanged data size), because different limits can be reached depending on them.…”

“…Hence, depending on the configuration of request Type and Size parameters, the value of the Intensity of requests has to be chosen properly. The measures of the Intensity parameter is defined in [22] as the percentage of the maximum system's capacity. For High level we use the 80% of the limit, while the 10% of the limit is used for the Low level.…”

“…The duration time of each experiment was set to 24 hours (for a total amount of 192 hours), since we observed a memory depletion trend within that time in the experiment with the least stressful workload (i.e., exp 5). Therefore, by assuming that the more intensive workload, the higher aging trends are [22], 24 hours suffice to observe aging effects in all the experiments.…”