Changes in proper system functioning are often connected with violation of accessibility to its services. The goal of automatic detection of the functioning parameters changes leads to the necessity of the statistically distributed process imbalance presence to be estimated. The use of classic method of searching for the maximum of statistical process as the imbalance ( ) moment determination ( ) is suggested in this article. The novelty of such an approach is in building of the statistical process on the basis of statistics, which are typical for similarity criteria of polynomial schemes.
Basing on the suggested statistics the observation process trajectory is built, where the top value is marked. Time position of the marked top value is considered as the estimation of imbalance moment. The analysis of distribution process dependence in every observation moment, which is mentioned in this article, shows that in case of imbalance, the size of trusted interval makes up a value of about , where t is the volume of observations, if the imbalance is situated in an area, which is proportional to t, whether the t is rather large. During the analysis such a restriction is being modeled with the help of condition . It seems impossible to evaluate more accurate view of trusted interval, while the distributions of the process before and after imbalance are unknown.
The results of statistical processing among the imbalanced processes, which model a computer system workflow, while being infected by a virus, are represented in this article. The final conclusion on the efficiency of such a method is suggested to be based on experimental data received from the real observation processes.