Reliability, availability and maintainability analysis of hot standby database systems

The prominent objective of the present study is to assess the reliability, availability and maintainability of a hot standby database system by analyzing the failure data of the system. This study is also helpful in predicting the maintenance schedules and adopting effective maintenance strategies based on repair data recorded for the thirty months of a Database system. The reliability, availability, and maintainability of the system have been evaluated with the help of basic probability arguments. The parameters of well-known probability distributions have been estimated. The statistical analysis, trend and serial correlation, reliability, and maintainability analysis of the primary unit and hot standby unit are performed. It is observed that the database system was available for use at approximately 99.9997628% time. It is observed that no serial and trend correlation present in the failure and repair data. The best fit for the time between failures (TBF) of primary and hot standby units is Weibull Distribution having Anderson-Darling (AD) values 0.486 and 1.161 respectively. Though, best fitted distribution for the TTR of hot standby unit is 3-Parameter Loglogistic Distribution having AD value 1.112 while for primary unit is 3-Parameter Weibull distribution with AD value 1.004. Furthermore, as RAM analysis has been done to increase database system performance, so database designers and managers can use the findings of the present study.