Logean X, Dietrich F, Koppenhoefer S. Run-time monitoring of distributed applications. Proceedings of Middleware`98, The Lake District, England, 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....within an entity, our framework is capable to employ the logical clock technique and restore order. Nevertheless, because of the shared memory approach, the causality relation in the execution model is broken and a separate solution is sought in order to track causality in the system. In [15], Logean describes a method for run time monitoring of distributed applications that supports a software development process. His approach employs ordering technique deriving from logical clocks. The level of entity granularity is the CORBA object. Our framework enhances this approach with ....

Logean, X., F. Dietrich, H. Karamyan, S. Koppenhoefer. Run-time monitoring of distributed applications. Middleware'98, IFIP Intl. Conf. on Distributed Systems Platforms and Open Distributed Processing, Springer-Verlag, London, 2000, 459-473.

....As a part of the ORB, the monitoring support ensures order of the monitoring records (events) inside the ORB, however, the monitoring support does not offer consistency of the event model with respect to the events generated from a large distributed system that runs several ORB instances. In [8], a method is described for run time monitoring of distributed applications that supports a software development process. The approach provides consistent event model. The level of granularity is the CORBA object, whereas our monitoring framework enhances this approach with introduction of the ....

....that supports a software development process. The approach provides consistent event model. The level of granularity is the CORBA object, whereas our monitoring framework enhances this approach with introduction of the component abstraction. The configuration of the 8 monitoring code in [8] can be dynamic and supported by tools. By applying a notification service, our approach introduces an additional scheme for decoupling event sources from event consumers. Additionally, our monitoring framework supports several event types and support for different ORB implementations. 6. ....

Logean, X., Dietrich, F., Karamyan, H. and Koppenhoefer, S., "Run-time Monitoring of Distributed Applications", Proceedings of the IFIP International Conference on Distributed Systems Platforms and Open Distributed Processing (Middleware'98), pp. 459-473, Hudson River Valley, New York, USA, 3-7 April 2000.

....processes within an entity, our framework is capable to employ the logical clock technique and restore order. Nevertheless, because of the shared memory approach the causality relation in the execution model is broken and a separate solution is sought in order to track causality in the system. In [15] Logean describes a method for run time monitoring of distributed applications that supports a software development process. His approach employs ordering technique deriving from logical clocks. The level of entity granularity is the CORBA object. Our monitoring framework enhances this approach ....

Logean, X., Dietrich, F., Karamyan, H. and Koppenhoefer, S., "Run-time Monitoring of Distributed Applications", Proceedings of the IFIP International Conference on Distributed Systems Platforms and Open Distributed Processing (Middleware'98), pp. 459-473, Hudson River Valley, New York, USA, 3-7 April 2000.

....the specification of the properties. The properties specified with MOTEL are then constantly checked while the final service implementation is observed at run time and, if a property is violated by the service, an error message is given to the user. A detailed description of MOTEL can be found in [15]and[16] 3. A FORMAL MODEL FOR OODS In this section we will define, step by step, our model for object oriented distributed systems (OODS) and identify the observable events that are useful for the specification of behavioral properties. Figure 2 shows how the definitions introduced in this ....

....can be used to formally specify properties, to observe the behavior of the system, and to check at run time whether the specified properties are respected by the final implementation. It is outside the scope of this paper to describe in detail how MOTEL works; the interested reader is referred to [15]. The main features of MOTEL are summarized in the following. Specification of properties. Using the graphical interface of MOTEL, it is possible to compose properties in two ways: Using patterns: the user can select a pattern and then, guided by the tool, provides detailed information ....

Logean X, Dietrich F, Koppenhoefer S. Run-time monitoring of distributed applications. Proceedings of Middleware`98, The Lake District, England, 1998.

....of the properties. The properties specified with the guidance of MOTEL are then constantly checked while the final service implementation is observed at run time and, if a property is violated by the service, an error message is given to the user. A detailed description of MOTEL can be found in [32]. 3 A formal model for OODS In this section we will define our proposed model for object oriented distributed systems (OODS) This section is divided into subsections covering the notions of object class, thread process and OODS. We will define, step by step, our model and identify the ....

....properties are respected by the final implementation. As the name MOTEL indicates, it is primarily used to monitor and test. The main features of MOTEL are summarized in the following. It is outside the scope of this paper to describe in detail how MOTEL works; the interested reader is referred to [32]. ffl Specification of properties. Using the graphical interface of MOTEL, it is possible to compose properties in two ways: Using patterns: The user can select a pattern and then, guided by the tool, provides detailed information about the events involved in the property. At the current ....

X. Logean, F. Dietrich, and S. Koppenhoefer. Runtime monitoring of distributed applications. In Proceedings of Middleware`98, Lake District, England, September 1998.

....with the requests. The size of the timestamp is equivalent to the number of objects in the system. As the number of object in the system evolutes (depending on the creation and deletion of objects) our algorithm is capable of maintaining a dynamically sized timestamp. Our algorithm is detailed in [13]. C. Testing In order to be checked at run time, the LTL properties have to be translated in some machine form. Different approaches and tools are available for translating LTL formula into FSMs. In [14] Dillon developed a method to translate temporal logic formula into test oracles. The method ....

X. Logean, F Dietrich, S Koppenhoefer, and J.-P. Hubaux, "Runtime monitoring of distributed applications," Tech. Rep. available from the authors, Swiss Federal Institute of Technology, Lausanne, Communications System Division, 1998.

....of formal methods in the industry if their specific needs are not ignored. In this section, we show how our proposal addresses the industrial concerns discussed in Section 4. It is outside the scope of this paper to give a detailed technical description that can be found elsewhere [25] 61] [63] [24] 62] Technical details are only provided as far as it it necessary for our discussion of how and why our proposal caught the interest of the industry and is being integrated in an industrial platform. In our project we focused on the validation of object oriented distributed systems in ....

....a MOnitoring and TEsting tooL (MOTEL) that we have developed. MOTEL will automatically map the model properties to implementation properties and check at service run time whether the implemented system is violating the properties expressed on our model. A description of MOTEL can be found in [63]. We express behavioral constraints independently of a specific implementation language by relying completely on a set of observable events. The idea of event based behavioral abstraction has been frequently used (e.g. for testing [26] and debugging [3] and is especially useful for ....

[Article contains additional citation context not shown here]

X. Logean, F. Dietrich, and S. Koppenhoefer. Run-time monitoring of distributed applications. In Proceedings of Middleware`98, Lake District, England, September 1998.

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