Measuring Unreliability and Unavailability using COMPASS
Hello, I have been trying out COMPASS for the past few weeks, but I have not quite managed to get what I want out of it. I am mostly interested in quantitative measures such as unreliability and unavailability. Using the battery_sensor example provided with COMPASS, I could at least obtain a measure for unreliability in the Safety/Fault Tree Evaluation tab. However, this feature works by generating fault trees - I have been under the impression that COMPASS has more expressive power than fault trees. And I have no idea how to measure unavailability using COMPASS - the battery_sensor example is not repairable and neither are most of the examples provided with COMPASS. I believe the Performability/IMC Analysis feature should also be able to give me quantitative results, but all I can get out of it is "Long Run Average: 0.0". I have noticed the warning "The root component contains ports. Performability results are undefined.", but even with a SLIM model I made myself featuring no ports in its root component, I get the same result. I am using a property of the form "Globally, {not sys.is_alive} holds eventually with probability > 0.0". Is there any way to obtain measures of unreliability and more importantly, unavailability from COMPASS? Is the only way to get these measures via Fault Tree Generation? Additionally, I have been wondering if there is any kind of archive of this mailing list - I have been looking for one but have not found any. Best Regards, Hermann Walth
Dear Hermann, thank you for your interest in COMPASS. Indeed, there are some issues with probabilistic verification and performability analysis. This is explained in Section 6 of the COMPASS tutorial, where a variant of the battery_sensor model is presented and probabilistic analyses are exemplified. Someone from RWTH Aachen University -- who implemented this part -- can perhaps follow up with some more details. As regards this mailing list, it is a closed support list, so there is no public archive of it. Best regards, Marco On 09-Jul-21 15:44, Walth, Hermann wrote:
Hello,
I have been trying out COMPASS for the past few weeks, but I have not quite managed to get what I want out of it. I am mostly interested in quantitative measures such as unreliability and unavailability.
Using the battery_sensor example provided with COMPASS, I could at least obtain a measure for unreliability in the Safety/Fault Tree Evaluation tab. However, this feature works by generating fault trees - I have been under the impression that COMPASS has more expressive power than fault trees. And I have no idea how to measure unavailability using COMPASS - the battery_sensor example is not repairable and neither are most of the examples provided with COMPASS.
I believe the Performability/IMC Analysis feature should also be able to give me quantitative results, but all I can get out of it is "Long Run Average: 0.0". I have noticed the warning "The root component contains ports. Performability results are undefined.", but even with a SLIM model I made myself featuring no ports in its root component, I get the same result. I am using a property of the form "Globally, {not sys.is_alive} holds eventually with probability > 0.0".
Is there any way to obtain measures of unreliability and more importantly, unavailability from COMPASS? Is the only way to get these measures via Fault Tree Generation?
Additionally, I have been wondering if there is any kind of archive of this mailing list - I have been looking for one but have not found any.
Best Regards,
Hermann Walth
_______________________________________________ COMPASS-support mailing list -- compass-support@lists.rwth-aachen.de To unsubscribe send an email to compass-support-leave@lists.rwth-aachen.de https://lists.rwth-aachen.de/postorius/lists/compass-support.lists.rwth-aach...
============================================================================== Marco Bozzano, PhD, Researcher at FBK Via Sommarive 18, Povo, 38123 Trento (Italy) ph.: +39-0461-314367 mobile: +39-366-6722580 (calls only, no SMS) bozzano@fbk.eu, http://es.fbk.eu/people/bozzano/ ============================================================================== -- -- Le informazioni contenute nella presente comunicazione sono di natura privata e come tali sono da considerarsi riservate ed indirizzate esclusivamente ai destinatari indicati e per le finalità strettamente legate al relativo contenuto. Se avete ricevuto questo messaggio per errore, vi preghiamo di eliminarlo e di inviare una comunicazione all’indirizzo e-mail del mittente. -- The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. If you received this in error, please contact the sender and delete the material.
Hello Again, Thank you for pointing me towards the specific section in the documentation that talks about my issues. Using that, I have managed to measure the unreliability of a system. Unfortunately, I am still in the dark as to how I would measure unavailability – my first attempt involved setting a generic property to “recurrence” rather than “existence”, but I saw the IMC analysis is not even available for such a property. Apart from this, I would like to know how error propagation works. The documentation I have read explains how to declare error propagation ports in an error model, but not how to connect such ports between error models. How would I make an error model react to an error that is being propagated from another error model? Best Regards, Hermann From: Marco Bozzano<mailto:bozzano@fbk.eu> Sent: 12 July 2021 07:45 To: Walth, Hermann<mailto:hermann.walth@rwth-aachen.de>; COMPASS Consortium<mailto:compass-support@lists.rwth-aachen.de> Cc: Khan, Shahid<mailto:shahid.khan@cs.rwth-aachen.de> Subject: Re: [COMPASS-SUPPORT] Measuring Unreliability and Unavailability using COMPASS Dear Hermann, thank you for your interest in COMPASS. Indeed, there are some issues with probabilistic verification and performability analysis. This is explained in Section 6 of the COMPASS tutorial, where a variant of the battery_sensor model is presented and probabilistic analyses are exemplified. Someone from RWTH Aachen University -- who implemented this part -- can perhaps follow up with some more details. As regards this mailing list, it is a closed support list, so there is no public archive of it. Best regards, Marco On 09-Jul-21 15:44, Walth, Hermann wrote:
Hello,
I have been trying out COMPASS for the past few weeks, but I have not quite managed to get what I want out of it. I am mostly interested in quantitative measures such as unreliability and unavailability.
Using the battery_sensor example provided with COMPASS, I could at least obtain a measure for unreliability in the Safety/Fault Tree Evaluation tab. However, this feature works by generating fault trees - I have been under the impression that COMPASS has more expressive power than fault trees. And I have no idea how to measure unavailability using COMPASS - the battery_sensor example is not repairable and neither are most of the examples provided with COMPASS.
I believe the Performability/IMC Analysis feature should also be able to give me quantitative results, but all I can get out of it is "Long Run Average: 0.0". I have noticed the warning "The root component contains ports. Performability results are undefined.", but even with a SLIM model I made myself featuring no ports in its root component, I get the same result. I am using a property of the form "Globally, {not sys.is_alive} holds eventually with probability > 0.0".
Is there any way to obtain measures of unreliability and more importantly, unavailability from COMPASS? Is the only way to get these measures via Fault Tree Generation?
Additionally, I have been wondering if there is any kind of archive of this mailing list - I have been looking for one but have not found any.
Best Regards,
Hermann Walth
_______________________________________________ COMPASS-support mailing list -- compass-support@lists.rwth-aachen.de To unsubscribe send an email to compass-support-leave@lists.rwth-aachen.de https://lists.rwth-aachen.de/postorius/lists/compass-support.lists.rwth-aach...
============================================================================== Marco Bozzano, PhD, Researcher at FBK Via Sommarive 18, Povo, 38123 Trento (Italy) ph.: +39-0461-314367 mobile: +39-366-6722580 (calls only, no SMS) bozzano@fbk.eu, http://es.fbk.eu/people/bozzano/ ============================================================================== -- -- Le informazioni contenute nella presente comunicazione sono di natura privata e come tali sono da considerarsi riservate ed indirizzate esclusivamente ai destinatari indicati e per le finalità strettamente legate al relativo contenuto. Se avete ricevuto questo messaggio per errore, vi preghiamo di eliminarlo e di inviare una comunicazione all’indirizzo e-mail del mittente. -- The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. If you received this in error, please contact the sender and delete the material.
participants (2)
-
Marco Bozzano -
Walth, Hermann