Ett sätt att formellt beskriva och mäta hälsotillståndet på en integrationsplattform

Sammanfattning: An integration platform is a platform that is used to integrate data between two incompatible systems. The architecture is fully adapted to the customers’ requirements and therefore looks different between systems. In order to be able to describe and judge the state of the health of an integration platform, there are many factors to consider. This research has been designed to find out which general factors that should be considered in the assessment and description of the health status of an integration platform. A thorough and detailed research and analysis has been made about the general factors that previous researches has highlighted as important. After that, a big analysis of the detected problems was made in comparison to the general factors to draw parallels, conclusions and common denominators if the detected errors depend on the general factors or not. The goal of this was to confirm the validity of the previous research and see if any new factors appear in this integration platform. Two prototypes have been produced to present an overall picture of the actual health status of the integration platform. These two prototypes are developed using assessment in the systems that using the integration platform. These two prototypes have used several factors and parameters in their assessment. A baseline was developed to represent a value that estimate the normal state of the investigated platform. The estimated baseline is an estimate of the normal state within which response time is assumed to be working. This was made to have a value to compare both current and historical health status in the integration platform. Percentiles were also developed to produce an average value. The 50th percentile was used as a general average value and the 90th percentile is used to illustrate and detect severe deviations. The result of the two prototypes is, respectively, an assessment of the actual health status of the integration platform. For more correctness and more detailed assessment of the actual health state, each subsystem in the researched system should be assessed individually because different subsystems affect each other and the health status different. This can lead to misleading picture of the actual health state of the integration platform. This research has combined all the subsystems in the research system in order to be able to describe and measure the actual state of the integration platform.