Performance Monitoring
Keeping under control your Ground Segment facilities.
A Ground Segment consists of several complex facilities interacting among each other. The data production and exchange between the facilities is in most of the cases driving the behaviour of the system, and a failure in a given step may compromise the overall performance, resulting in loss of data.
In many cases the data exchange is materialised as electronic files that are disseminated using a network that communicates the various components.
DEIMOS Space has participated actively with ESA in the development of dedicated facilities used for monitoring the performance of the Ground Segments. It detects anomalies in the instrument processors, product archiving, and product dissemination. It monitors the product generation dataflow from the acquisition to the processing and to the dissemination to the user services, and is also capable of comparing this information with the original planning made for the mission.
To do this, it gets the input and output data from the GS facilities. It calculates the expected outputs according to a model of each facility and compares them to the actual outputs for obtaining the deviation (anomalies).
Based on the detected problems, anomalies are raised to inform the operator. These anomalies can also re-processing or re-dissemination requests in order to close the loop.
All current ESA Earth Explorer missions (CryoSat, GOCE, SMOS and Aeolus) base this analysis on tools implemented by DEIMOS. Moreover, the approach has been to use a generic mission-independent core that can be customised for each individual mission. The core system is the QUARC infrastructure (Quality Analysis and Reporting Computer) originally developed for ENVISAT but designed as an independent system able to support any mission.
At the core of the QUARC lies a Database, which keeps the monitoring parameters into a common data model. This generic, but highly flexible and adaptable, representation is the key in monitoring approach. This long-term storage allows users analyzing the monitoring data in any past period (i.e. day, week, month…).
All “measurements” taken by the system on the real world (e.g. files produced by a given facility, reports generated by another one) are transformed to parameters of the common data model and then ingested into the database.
Once the data is available, a powerful high-level language is able to manipulate the data, performing all kind of operations and correlations to derive the monitoring results. Modelling of the GS facilities, to identify the expected behaviour is done at this step. The result of the monitoring task is formatted as HTML reports that are either published in a web server of sent by e-mail to the mission responsible. Interactive analysis of the data is also possible, either by re-uting monitoring tasks or by exporting the data to a Gantt Tool application.
Retrieval and dissemination of input / output data is made using a highly reliable and configurable subsystem that can also be reused in the frame of other applications.
|
