A. Aghasaryaiu, E. Fabre, A. Benveniste, R. Boubour, and C. Jard, `A petri net approach to fault detection and diagnosis in distributed systems', in Proc. 1997.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... [4] 5] Decentralized and asynchronous diagnosis in DES has been addressed more recently (see, for example, 6] 7] 8] Also related to the work presented in this paper are diagnosis methodologies that identify faults in distributed settings using partially stochastic networks [9] [10], 11] 12] The approach in [13] develops template monitoring schemes based on concurrent observation of events and applies them to fault monitoring in automated manufacturing systems. In [14] the behavior of a DES is represented by a Petri net and techniques are presented to reconstruct the ....

A. Aghasaryaiu, E. Fabre, A. Benveniste, R. Boubour, and C. Jard, "A Petri net approach to fault detection and diagnosis in distributed systems (Part II)," in Proc. of the 36th IEEE Conf. on Decision and Control, pp. 726--731, 1997.

.... [4] 5] Decentralized and asynchronous diagnosis in DES has been addressed more recently (see, for example, 6] 7] 8] Also related to the work presented in this paper are diagnosis methodologies that identify faults in distributed settings using partially stochastic networks [9], 10] 11] 12] The approach in [13] develops template monitoring schemes based on concurrent observation of events and applies them to fault monitoring in automated manufacturing systems. In [14] the behavior of a DES is represented by a Petri net and techniques are presented to reconstruct ....

A. Aghasaryaiu, E. Fabre, A. Benveniste, R. Boubour, and C. Jard, "A Petri net approach to fault detection and diagnosis in distributed systems (Part I)," in Proc. of the 36th IEEE Conf. on Decision and Control, pp. 720--725, 1997.

....insight into the role of communication in diagnostic design for distributed systems. The relevant literature is as follows. Fault detection in distributed discrete event systems has been investigated in the context of Petri Net models with applications to telecommunication networks in [2] and [1]. Similar problems have been studied in the context of template languages for fault monitoring in [9] Within Research supported in part by Office of Naval Research grant 442427 25828 and CALTRANS PATH MOU 331. the context of automata and language theoretic models, centralized diagnostic design ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, and C. Jard. A petri net approach to fault detection and diagnosis in distributed systems (part 2). Proc. 36th IEEE CDC, December 1997.

....alarm events is modelled in some way or another. The articles [6] 7] 8] belongs to this family. The authors formulate diagnosis as an optimisation problem, this has some relation with the present approach. The case of event correlation in network management also motivated the series of papers [9][10][11] on which the present paper relies. Diagnosis architectures. In this paper, motivated by the example of telecommunication network diagnosis and management, we focus on diagnosis algorithms that are suitable to truly distributed, asynchronous, systems. Figure 1 illustrates our purpose by ....

....meaning that its behaviour changes dynamically, due to reconfigurations. In this paper we focus on the third case. Modelling approaches. In this paper we follow an approach based on partial order models of time and making no use of a global notion of state. This approach was introduced in [9][10][11] and is motivated by the figures 2 and 3. Figure 2 depicts a typical fault propagation in a SDH SONET network management system. The different boxes are network elements, and the different links correspond to the different layers in the SDH SONET hierarchy. The fault shown is a fault in STM1 ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard. A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets. CDC'97 Proceedings, San Diego, December 1997.

....related by constraints [4] The interested reader will find there the definition of a generalized likelihood, the counterpart of the HMM (Hidden Markov Model) paradigm, and algorithms to solve MAP (Maximum a Posteriori) problems. PSPNs have been originally described with this formalism [1] 2] [3]. For lack of space, we prefer to present only its specialization to Petri nets and refer the reader to the above mentioned references for a broader view on the subject. 3.3.1 Attributes of a place As usual SPNs, a PSPN is obtained as a Petri net that handles not only tokens but also extra ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard, "A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets." CDC'97 Proceedings, San Diego, December 1997.

....global supervisor. Protocols defining the nature of alarms incorporate various mechanisms that allow to keep track of causal or temporal dependencies between them 1 . We will not enter into the details of each protocol in this paper (an example concerning the SDH protocol can be found in [2] and [1]) However, a reasonable and rather general model states that an alarm bears 1 information on the fault that generated it, and 2 information of the kind has been caused by previous alarms [list of alarms] or 2 information like appeared necessarily after alarms [list of alarms] Of course, ....

....variables, related by constraints [4] The interested reader will find there the definition of a generalized likelihood, the counterpart of the HMM (Hidden Markov Model) paradigm, and algorithms to solve MAP (Maximum a Posteriori) problems. PSPNs have been originally described with this formalism [1], 2] 3] For lack of space, we prefer to present only its specialization to Petri nets and refer the reader to the above mentioned references for a broader view on the subject. 3.3.1 Attributes of a place As usual SPNs, a PSPN is obtained as a Petri net that handles not only tokens but also ....

A. Aghasaryan, R. Boubour, E. Fabre, C. Jard, A. Benveniste, "A Petri net approach to fault detection and diagnosis in distributed systems." IRISA Research Report no. 1117, August 1997.

....the diagnosis algorithm we obtain is made of completely asynchronous processes, and does not require a centralized supervisor, nor a centralized data structure. The distributed Viterbi algorithm we propose in this paper is dedicated to a stochastic distributed system (it extends some results of [6, 7, 8], developed for Petri nets) Special semantics are designed for the trajectories of the system, in order to represent explicitly the concurrency of events in subsystems. These semantics are the key to reduce the curse of dimensionality, because it is precisely the amount of concurrency in a ....

....partial order semantics, trajectories are partial orders ef events instead of sequences. They are represented by a directed acyclic graph (DAG) of events, which is usually chosen to be the transitive reduction of the partial order (see the notion of occurence net in [3] the notion of puzzle in [6, 8], or, more generally, the notion of unfolding in [4] S A S C S B Figure 2: Partial order semantics for a trajectory of SAkSC kSB . White blocks indicate transitions of TA , grey blocks those of TB . Small blocks represent local transitions (that operate on local states only) whereas ....

[Article contains additional citation context not shown here]

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard, "A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets." CDC'97 Proceedings, San Diego, December 1997.

....alarm events is modelled in some way or another. The articles [7] 8] 20] belongs to this family. The authors formulate diagnosis as an optimisation problem, this has some relation with the present approach. The case of event correlation in network management also motivated the series of papers [6][2][3] on which the present paper relies. Irisa Asynchronous diagnosis 5 In this paper, motivated by the example of telecommunication network diagnosis and management, we focus on diagnosis algorithms that are suitable to truly asynchronous systems. Figure 1 illustrates our purpose by showing two ....

....networks, which requires considering distributed supervisors, is discussed in the full paper [1] In this paper we focus on the second case of figure 1. We follow an approach based on partial order models of time and making no use of a global notion of state. This approach was introduced in [6][2][3] and is motivated by the figures 2 and 3. Figure 2 depicts a typical fault propagation in a SDH SONET network management system. The different boxes are network elements, and the different links correspond to the different layers in the SDH SONET hierarchy. The fault shown is a fault in STM1 ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard. A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets. CDC'97 Proceedings, San Diego, Dec. 1997.

....behaviors) has not been developed. Inspired by statistical mechanics and random field theory, Benveniste, L evy and Fabre have proposed more recently [4] a theory of interacting finite systems of random variables, but again extension to stochastic processes with dynamics was not available. In [5][1][2] Fabre et al. have considered for the first time a new class of stochastic Petri nets in which interleaving of concurrent transitions was not randomized, we called them Partially Stochastic Petri nets (Pspn) In this reference, an in depth comparison of this new model with existing stochastic ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard. A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets. CDC'97 Proceedings, San Diego, December 1997.

....network system: the SDH network (Synchronous Data Hierarchy) in collaboration with CNET (the research center of France Telecom) The model and algorithm presented here serve also as a basis for a probabilistic approach to fault detection. This study is conducted in an other team of our laboratory [1]. The technical contribution can be announced in three points: 1. a new way of modeling alarm correlation by focusing on causal dependencies between faults and alarms, 2. a dedicated detection algorithm based on an on line unfolding of Petri nets, 3. making appear good properties of the model ....

....The use of time stamp mechanism prevents from noise due to physical clock distortion of observation. This model is illustrated with an example of SDH network. The detection algorithm is programmed with the Eiffel language, and we plan real experiments. A joined work with the use of probabilities [1] completes this approach, in order to propose a new complete way to diagnose faults in telecommunication network. A short term research is to define precisely the way to provide diagnosis (i.e. explanation) and to be able to distribute the supervising activity on sensors: 1. A diagnostic could ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, and C. Jard. A Petri Net Approach to Fault Detection and Diagnosis in Distributed Systems. part II: Extending Viterbi Algorithm and Hmm Techniques to Petri Nets. To appear in IEEE Conference on Decision and Control, december 1997.

....are partial orders ef events instead of sequences, so that concurrent events are not ordered in time. They are represented by a directed acyclic graph (DAG) of events, usually taken to be the transitive reduction of the partial order (see the notions of occurence net in [3] of puzzle in [6, 8], or of unfolding in [4] A partial order of events of S (also called an event graph) is a valid trajectory for S i# its restriction to events of SA (resp. SB , SC )isa valid sequence of events for SA (resp. SB , SC ) and if the partial order of events coincides with the ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard, "A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets." CDC'97 Proceedings, San Diego, December 1997.

....to interacting stochastic processes has not been developed. Inspired by statistical mechanics and random field theory, we have proposed more recently [4] a theory of interacting finite systems of random variables, but again extension to stochastic processes with dynamics was not available. In [5, 6, 7] we have considered for the first time a new class of stochastic Petri nets in which interleaving of concurrent transitions was not randomized, we called them Partially Stochastic Petri nets (Pspn) In this reference, an in depth comparison of this new model with existing stochastic extensions of ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour, C. Jard. A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and HMM techniques to Petri nets. CDC'97 Proceedings, San Diego, December 1997.

....work concurrently; ffl But they may sometimes interact, reflecting fault propagation among components; ffl Fault models for each component can be triggered and controlled by the discrete controller for the system operating modes. Based on [9, 7] such a natural framework has been proposed in [10, 2, 1], where a new class of so called partially stochastic Petri nets (pspn) is proposed, together with an extension of classical hmm techniques to this class of nets. It is out of the scope of the present paper to describe this new modular framework and discuss its relevance for taking into account a ....

A. Aghasaryan, E. Fabre, A. Benveniste, R. Boubour and C. Jard (1997). A Petri net approach to fault detection and diagnosis in distributed systems. Part II : extending Viterbi algorithm and hmm techniques to Petri nets. Proc. IEEE CDC'97, San Diego, CA.

No context found.

A. Aghasaryaiu, E. Fabre, A. Benveniste, R. Boubour, and C. Jard, `A petri net approach to fault detection and diagnosis in distributed systems', in Proc. 1997.

No context found.

A. Aghasaryaiu, E. Fabre, A. Benveniste, R. Boubour, and C. Jard, `A petri net approach to fault detection and diagnosis in distributed systems', in Proc. 1997.

No context found.

A. Aghasaryan, E. Fabre, A. Benveniste, and R. Boubour, "A Petri net approach to fault detection and diagnosis in distributed systems (Part 2),", IEEE Conference on Decision and Control, pp. 726--731, San Diego, CA, 1997.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC