A comprehensive metric to evaluating software aging

Software aging refers to the phenomenon that, after a long-time running, a software system may show a degraded performance, or an increased occurrence rate of hang/crash failures or occupy overmuch resources of computer system. Software aging causes huge loss in safety/business-critical Web applications. Response time and failure rate are important metrics to identify software aging and/or evaluating the extent of aging. Two-dimensional metric is accurate, but not clearer than one-dimensional metric. Furthermore, both metrics are correlated. In this paper, we construct the relationship between both metrics with linear regression method, and get a resulted one-dimensional metric to evaluate software aging and evolving process comprehensively.