Minimizing Safety Risks in Complex Systems

There is a need to develop an adaptive model by which the changes in the testing and upgrade processes for complex software can be observed.The metrics obtained from a software system and a sufficient data base for software failures and operating hours, help to develop a model for predicting the reliability-relevant failures based on actual data and fitting a distribution. However, the collection, identification and classification of the failure data is not timely enough to identify the cause of the failure.From the system engineering perspective testing to an acceptable level is needed to validate the system capability and performance. Testing in a perceived operational environment reduces the risk of the system not meeting the required capabilities. Historically this risk was evaluated in terms of cost, schedule and data. This paper will examine the safety hazards to system design by examining the consequences based on the interaction of data and function and the ensuing behavior complicated by an uncertain input from the human. The paper will identify and develop metrics and interactions thereof which when varied can change the test profile to minimize the uncertainty or risk of the system performance in realistic scenarios for enterprise service systems minimizing safety risk.