Issue 18567: Questions on observators within GQAM and SAM (marte-rtf) Source: Commissariat a l Energie Atomique-CEA (Dr. Sebastien Gerard, sebastien.gerard(at)cea.fr) Nature: Uncategorized Issue Severity: Summary: I was wondering, why two kinds of observators was designed within the GQAM as denoted in the figure shown below: Why not gathering both into one single stereotype? I propose to keep the second one, GaLatencyObs. In addition, I was also wondering why the <<SaSchedObs>> stereotype defined within SAM was not inheriting from <<GaLatencyObs>> instead of <<GaTimedObs>>? What do you think? Resolution: Revised Text: Actions taken: March 19, 2013: received issue Discussion: End of Annotations:===== m: GERARD Sebastien 166342 To: "marte-rtf@omg.org" CC: TUCCI Sara Subject: Questios on observators within GQAM and SAM Thread-Topic: Questios on observators within GQAM and SAM Thread-Index: Ac4k13O8xTqVBAbtTXOmfZu7Tzg8VQ== Date: Tue, 19 Mar 2013 19:26:41 +0000 Accept-Language: fr-FR, en-US X-MS-Has-Attach: yes X-MS-TNEF-Correlator: x-originating-ip: [132.166.88.105] x-tm-as-product-ver: SMEX-10.2.0.1135-6.800.1017-19728.003 x-tm-as-result: No--50.927600-0.000000-31 x-tm-as-user-approved-sender: Yes x-tm-as-user-blocked-sender: No X-Virus-Scanned: amavisd-new at omg.org X-Brightmail-Tracker: AAAAAR0Z04M= X-Brightmail-Tracker: AAAAAA== Hi all, I was wondering, why two kinds of observators was designed within the GQAM as denoted in the figure shown below: Why not gathering both into one single stereotype? I propose to keep the second one, GaLatencyObs. In addition, I was also wondering why the <> stereotype defined within SAM was not inheriting from <> instead of <>? What do you think? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus From: GERARD Sebastien 166342 To: Julio Medina , "marte-rtf@omg.org" , "Espinoza Ortiz, Huascar" CC: TUCCI Sara Subject: RE: Questios on observators within GQAM and SAM Thread-Topic: Questios on observators within GQAM and SAM Thread-Index: Ac4k13O8xTqVBAbtTXOmfZu7Tzg8VQBJt6cgAABIozAAAvPPgAAKVLHA Date: Thu, 21 Mar 2013 13:06:50 +0000 Accept-Language: fr-FR, en-US X-MS-Has-Attach: yes X-MS-TNEF-Correlator: x-originating-ip: [132.166.88.105] x-tm-as-product-ver: SMEX-10.2.0.1135-6.800.1017-19734.003 x-tm-as-result: No--70.612300-0.000000-31 x-tm-as-user-approved-sender: Yes x-tm-as-user-blocked-sender: No X-Virus-Scanned: amavisd-new at omg.org X-Brightmail-Tracker: AAAAAh0YqyAdGdOD X-Brightmail-Tracker: AAAAAA== Ok. Thanks. Aditional question o the meaning of the attributes of the GaLatencyObs. My understanding for latency and miss attributes is the following: · latency is an integer value used to specify the latency of the timing observation. In this case, the latency denotes the time between the start event reception and the actual start execution time of the triggered behavior, · miss is a real value denoting the miss ratio admitted with respect the observed end timing denoting by the attributes endObs, Do you confirm or no, I am right? For the maxJitter attribute, it is said in the spec that it .represents a maximum deviation with which a periodic internal event is generated.. I do not see to which periodic internal event it is made reference to ? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus De : Julio Medina [mailto:julio.medina@unican.es] Envoyé jeudi 21 mars 2013 10:08 À: GERARD Sebastien 166342 Cc : Espinoza Ortiz, Huascar; TUCCI Sara Objet : Re: Questios on observators within GQAM and SAM Hi Sé hi all, The reason for having SchedulingObserver inheriting from TimedObserver, instead of from LatencyObserver was basically that aspects included in SchedulingObserver were really not that visible at user level, but as data internal to the analysis process, and hence that stereotype was expected to be rarely used. These data were collected from some particular tools and techniques and are not really necessary as input data but may be interesting as output data. In any case, I think that changing the inheritance of <> from <> to <> is not harmful, though in general I will not advocate for making SAM stereotypes completely self-contained..... Julio El 21/03/2013 7:45, Espinoza Ortiz, Huascar escribióSorry I just realised that what you proposed was different. .why the <> stereotype defined within SAM was not inheriting from <> instead of <>. OK, in that case, I cannot see now any problem to do so. Huascar De: Espinoza Ortiz, Huascar Enviado el: jueves, 21 de marzo de 2013 7:41 Para: 'GERARD Sebastien 166342' CC: TUCCI Sara; Julio Medina (julio.medina@unican.es) Asunto: RE: Questios on observators within GQAM and SAM Regarding the first question I answered in my previous email to you Seb. Regarding the second one, it is because that.s an observer that look at other information (suspensions, blocking time, overlaps), instead of latency aspects. The reason why it is separated is because PAM needed GALatencyObs and not SASchedObs. Except for that you can merge both I guess. Cheers, PS: Now I.ll be traveling. Look at any message during the weekend. Huascar De: GERARD Sebastien 166342 [mailto:Sebastien.GERARD@cea.fr] Enviado el: martes, 19 de marzo de 2013 20:27 Para: marte-rtf@omg.org CC: TUCCI Sara Asunto: Questios on observators within GQAM and SAM Hi all, I was wondering, why two kinds of observators was designed within the GQAM as denoted in the figure shown below: Why not gathering both into one single stereotype? I propose to keep the second one, GaLatencyObs. In addition, I was also wondering why the <> stereotype defined within SAM was not inheriting from <> instead of <>? What do you think? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus X-Virus-Scanned: amavisd-new at unican.es Date: Thu, 21 Mar 2013 16:57:15 +0100 From: Julio Medina User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/20130307 Thunderbird/17.0.4 To: GERARD Sebastien 166342 CC: "marte-rtf@omg.org" , "Espinoza Ortiz, Huascar" , TUCCI Sara Subject: Re: Questios on observators within GQAM and SAM X-Originating-IP: [193.144.198.181] X-Virus-Scanned: amavisd-new at omg.org X-Brightmail-Tracker: AAAAAR0YqyA= X-Brightmail-Tracker: AAAAAA== Hi Sé hi everyone, The first two definitions you mention are not correct. In the context of LatencyObserver, the attribute Latency is kind of clearly described in Annex F: It specifies a duration observation between startObs and endObs TimedInstantObservations. It is NOT an integer but an NFP_Duration. This means that the startObs and endObs need to be indicated and usually refer to eventOcurrences in the end-to-end flow; for example one of the stimulus, and some of the precedence relations or the end of one of the steps/scenarios of the flow. The missRatio is used for soft timing constraints, it indicates the admitted(required) or actual(measured, calculated) percentage of the times in which the .required. latency are to be or have not been respected. In Schedulability analysis terms this is the ratio of missed deadlines. (This text may be probably used to clarify its definition, though it is made in the context of GQAM) The definition of maxJitter may be improved, since the reference to a "periodic event" may be confusing : maxJitter: NFP_Duration [*] Maximum deviation value of the latency. It represents a maximum deviation in the range of possible values for the observed latency. The output jitter is calculated as the difference between a worst-case latency time and the best-case latency time for the observed event (endObs) measured from a reference event (startObs). Regards, Julio El 21/03/2013 14:06, GERARD Sebastien 166342 escribióOk. Thanks. Aditional question o the meaning of the attributes of the GaLatencyObs. My understanding for latency and miss attributes is the following: · latency is an integer value used to specify the latency of the timing observation. In this case, the latency denotes the time between the start event reception and the actual start execution time of the triggered behavior, · miss is a real value denoting the miss ratio admitted with respect the observed end timing denoting by the attributes endObs, Do you confirm or no, I am right? For the maxJitter attribute, it is said in the spec that it .represents a maximum deviation with which a periodic internal event is generated.. I do not see to which periodic internal event it is made reference to ? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus De : Julio Medina [mailto:julio.medina@unican.es] Envoyé jeudi 21 mars 2013 10:08 À: GERARD Sebastien 166342 Cc : Espinoza Ortiz, Huascar; TUCCI Sara Objet : Re: Questios on observators within GQAM and SAM Hi Sé hi all, The reason for having SchedulingObserver inheriting from TimedObserver, instead of from LatencyObserver was basically that aspects included in SchedulingObserver were really not that visible at user level, but as data internal to the analysis process, and hence that stereotype was expected to be rarely used. These data were collected from some particular tools and techniques and are not really necessary as input data but may be interesting as output data. In any case, I think that changing the inheritance of <> from <> to <> is not harmful, though in general I will not advocate for making SAM stereotypes completely self-contained..... Julio El 21/03/2013 7:45, Espinoza Ortiz, Huascar escribióSorry I just realised that what you proposed was different. .why the <> stereotype defined within SAM was not inheriting from <> instead of <>. OK, in that case, I cannot see now any problem to do so. Huascar De: Espinoza Ortiz, Huascar Enviado el: jueves, 21 de marzo de 2013 7:41 Para: 'GERARD Sebastien 166342' CC: TUCCI Sara; Julio Medina (julio.medina@unican.es) Asunto: RE: Questios on observators within GQAM and SAM Regarding the first question I answered in my previous email to you Seb. Regarding the second one, it is because that.s an observer that look at other information (suspensions, blocking time, overlaps), instead of latency aspects. The reason why it is separated is because PAM needed GALatencyObs and not SASchedObs. Except for that you can merge both I guess. Cheers, PS: Now I.ll be traveling. Look at any message during the weekend. Huascar De: GERARD Sebastien 166342 [mailto:Sebastien.GERARD@cea.fr] Enviado el: martes, 19 de marzo de 2013 20:27 Para: marte-rtf@omg.org CC: TUCCI Sara Asunto: Questios on observators within GQAM and SAM Hi all, I was wondering, why two kinds of observators was designed within the GQAM as denoted in the figure shown below: Why not gathering both into one single stereotype? I propose to keep the second one, GaLatencyObs. In addition, I was also wondering why the <> stereotype defined within SAM was not inheriting from <> instead of <>? What do you think? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus From: GERARD Sebastien 166342 To: Julio Medina CC: "marte-rtf@omg.org" , "Espinoza Ortiz, Huascar" , TUCCI Sara Subject: RE: Questios on observators within GQAM and SAM Thread-Topic: Questios on observators within GQAM and SAM Thread-Index: Ac4k13O8xTqVBAbtTXOmfZu7Tzg8VQBJt6cgAABIozAAAvPPgAAKVLHAAAPyRIAADslrIA== Date: Thu, 21 Mar 2013 22:01:15 +0000 Accept-Language: fr-FR, en-US X-MS-Has-Attach: yes X-MS-TNEF-Correlator: x-originating-ip: [132.166.88.106] x-tm-as-product-ver: SMEX-10.2.0.1135-6.800.1017-19736.000 x-tm-as-result: No--70.222100-0.000000-31 x-tm-as-user-approved-sender: Yes x-tm-as-user-blocked-sender: No X-Virus-Scanned: amavisd-new at omg.org X-Brightmail-Tracker: AAAAAh0YqyAdGdOD X-Brightmail-Tracker: AAAAAA== Thanks Julio. It is clearer now. seb ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus De : Julio Medina [mailto:julio.medina@unican.es] Envoyé jeudi 21 mars 2013 16:57 À: GERARD Sebastien 166342 Cc : marte-rtf@omg.org; Espinoza Ortiz, Huascar; TUCCI Sara Objet : Re: Questios on observators within GQAM and SAM Hi Sé hi everyone, The first two definitions you mention are not correct. In the context of LatencyObserver, the attribute Latency is kind of clearly described in Annex F: It specifies a duration observation between startObs and endObs TimedInstantObservations. It is NOT an integer but an NFP_Duration. This means that the startObs and endObs need to be indicated and usually refer to eventOcurrences in the end-to-end flow; for example one of the stimulus, and some of the precedence relations or the end of one of the steps/scenarios of the flow. The missRatio is used for soft timing constraints, it indicates the admitted(required) or actual(measured, calculated) percentage of the times in which the .required. latency are to be or have not been respected. In Schedulability analysis terms this is the ratio of missed deadlines. (This text may be probably used to clarify its definition, though it is made in the context of GQAM) The definition of maxJitter may be improved, since the reference to a "periodic event" may be confusing : maxJitter: NFP_Duration [*] Maximum deviation value of the latency. It represents a maximum deviation in the range of possible values for the observed latency. The output jitter is calculated as the difference between a worst-case latency time and the best-case latency time for the observed event (endObs) measured from a reference event (startObs). Regards, Julio El 21/03/2013 14:06, GERARD Sebastien 166342 escribióOk. Thanks. Aditional question o the meaning of the attributes of the GaLatencyObs. My understanding for latency and miss attributes is the following: · latency is an integer value used to specify the latency of the timing observation. In this case, the latency denotes the time between the start event reception and the actual start execution time of the triggered behavior, · miss is a real value denoting the miss ratio admitted with respect the observed end timing denoting by the attributes endObs, Do you confirm or no, I am right? For the maxJitter attribute, it is said in the spec that it .represents a maximum deviation with which a periodic internal event is generated.. I do not see to which periodic internal event it is made reference to ? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus De : Julio Medina [mailto:julio.medina@unican.es] Envoyé jeudi 21 mars 2013 10:08 À: GERARD Sebastien 166342 Cc : Espinoza Ortiz, Huascar; TUCCI Sara Objet : Re: Questios on observators within GQAM and SAM Hi Sé hi all, The reason for having SchedulingObserver inheriting from TimedObserver, instead of from LatencyObserver was basically that aspects included in SchedulingObserver were really not that visible at user level, but as data internal to the analysis process, and hence that stereotype was expected to be rarely used. These data were collected from some particular tools and techniques and are not really necessary as input data but may be interesting as output data. In any case, I think that changing the inheritance of <> from <> to <> is not harmful, though in general I will not advocate for making SAM stereotypes completely self-contained..... Julio El 21/03/2013 7:45, Espinoza Ortiz, Huascar escribióSorry I just realised that what you proposed was different. .why the <> stereotype defined within SAM was not inheriting from <> instead of <>. OK, in that case, I cannot see now any problem to do so. Huascar De: Espinoza Ortiz, Huascar Enviado el: jueves, 21 de marzo de 2013 7:41 Para: 'GERARD Sebastien 166342' CC: TUCCI Sara; Julio Medina (julio.medina@unican.es) Asunto: RE: Questios on observators within GQAM and SAM Regarding the first question I answered in my previous email to you Seb. Regarding the second one, it is because that.s an observer that look at other information (suspensions, blocking time, overlaps), instead of latency aspects. The reason why it is separated is because PAM needed GALatencyObs and not SASchedObs. Except for that you can merge both I guess. Cheers, PS: Now I.ll be traveling. Look at any message during the weekend. Huascar De: GERARD Sebastien 166342 [mailto:Sebastien.GERARD@cea.fr] Enviado el: martes, 19 de marzo de 2013 20:27 Para: marte-rtf@omg.org CC: TUCCI Sara Asunto: Questios on observators within GQAM and SAM Hi all, I was wondering, why two kinds of observators was designed within the GQAM as denoted in the figure shown below: Why not gathering both into one single stereotype? I propose to keep the second one, GaLatencyObs. In addition, I was also wondering why the <> stereotype defined within SAM was not inheriting from <> instead of <>? What do you think? Thanks, Cheers. Sé ------------------------------------------------------------------------------------------------------------------------------------------------ Séstien Gérd +33 (0)1 69 08 58 24 / +33(0)6 88 20 00 47 CEA Saclay Nano-INNOV Institut CARNOT CEA LIST DILS/Laboratoire d.Ingéerie dirigépar les modès pour les Systès Embarqué(LISE), Point Courrier n°174 91 191 Gif sur Yvette CEDEX Description : PapyrusLogo_SmallFormatwww.eclipse.org/papyrus