Computerized elements of spacecraft (S/C) typically produce error and event indications as they perform their processing. Some indications are related to routine activities or less common events, while others indicate off-nominal conditions detected onboard such as single event upsets. Finally, some errors or events are produced due to software defects. These are the topic of this paper, and we analyze the nature and frequency of reported errors/events to investigate the question of whether such reports can be used as a proxy for detecting software errors. While some indications are intended to point to internal errors detectable by the software itself, many of the indications of software defects are not explicitly intended to serve that purpose. Further analysis is required to determine that an indication is as the result of a software defect rather than some other cause. Our analysis is based on error and event data from Magnetospheric Multiscale (MMS). MMS is a constellation of four satellites measuring characteristics of the Earth's magnetosphere. See the MMS Mission for a comprehensive description of the mission, satellites, and instruments. Each satellite has about twenty-five instruments/controllers (that is, sources of indications) including the Central Instrument Data Processor (CIDP). The CIDP coordinates activities for the payload deck on each MMS observatory and is the interface between the S/C processor and the payload deck for instrument commanding and instrument telemetry reporting. Of particular interest, the CIDP stores science data files in a specialized recorder separate from the primary S/C recorder and manages science data playback including freeing of science files for reuse. Many CIDP indications are generated in processing the science files. MMS has been operating for more than four years including commissioning, the primary mission, and the extended mission. Given the number of instruments/controllers in the payload, four observatories, and more than four years of operation, a large volume of recorded errors and events reported by the CIDP are available for analysis. These were examined to determine the relationship between the reports and software reliability. Many reports do not indicate software defects; however, some point to defects. This paper evaluates the potential for using such reports as a proxy for determining software reliability of onboard flight software (FSW).
Are Onboard Errors/Events a Useful Proxy for Software Reliability?
2020-03-01
4697364 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Springer Verlag | 2012
|Onboard Software Dynamic Architecture
Springer Verlag | 2012
|British Library Conference Proceedings | 1996
|British Library Online Contents | 1996
|Onboard Software Static Architecture
Springer Verlag | 2012
|