No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Application configuration in a client-server distributed system.
Abstract
This paper discusses the configuration of clients and servers in a distributed system. The question addressed is how does a client choose a appropriate server instance for invocation. It is assumed that a number of server instances are available in the distributed system, each instance having different characteristics that the client may wish to optimise. To examine the issues in this configuration and resource management problem the authors build an object model containing clients ad servers. Using this model they examine the whole decision making process through the lifecycle of object design, implementation, and instantiation in a distributed system
... At viewpoint of efficient reconfiguration of distributed resources, the Olan model by IMAG-LSR Laboratory [7] and the Configuration Management model by HewlettPackard Laboratory [8] have been studying. Though the Olan model defined the language for supporting the dynamic reconfiguration and the design of distributed application configured by heterogeneous components, they did not consider the component mobility and the duplicated objects. ...
The Distributed Object Group Framework(DOGF) we constructed supports the grouping of distributed objects that are required
for distributed application. From the DOGF, we manage distributed application as a logical single view by applying the concept
of object group, therefore the framework can provide distributed transparency for client’ request and binding service between/among
objects. The DOGF also has an adaptive structure for supporting given application’s features. For that reason, we adapted
the object registry policy, the security policy, the load balancing policy, and the real-time policy to our framework. For
using practically the DOGF on distributed environment, we developed the Defence System against Invading Enemy Planes(DSIEP)
as a simulation program that is the distributed application with real-time properties. From all of results obtained via the
DSIEP simulation on the DOGF, we verified whether the DOGF supports the adaptability of distributed services and the functionality
of group management.
... However, the researches above have not been supporting the object management service for multi-duplicated objects and the load balancing service for server systems. From a viewpoint of efficient reconfiguration of distributed resources, the Olan model by IMAG-LSR Laboratory[7] and the Configuration Management model by Hewlett-Packard Laboratory[8] have been studied. The Olan model defined the language for supporting the dynamic reconfiguration. ...
In this paper, we constructed the Distributed Object Group Framework(DOGF) which is a reconfigurable architecture supporting
dynamically adaptation of distributed services. We also developed the distributed application simulator to verify the adaptability
and reconfigurability of our framework. The DOGF provides appropriate grouping of distributed objects for executing given
distributed applications as the point of view of a single logical view system, and supports the adaptive distributed services
for dynamic reconfiguration of distributed application. In constructing procedure of the DOGF, we designed the distributed
services that consist of the objects supporting service, the load balancing service, and the real-time service. We also implemented
the managing technique for dynamic reconfiguring the distributed services. Finally, for verifying whether the DOGF can support
the dynamic reconfigurability of distributed services or not, we developed the Defence System against Invading Enemy Planes(DSIEP)
simulator as a practical use of distributed application with real-time property on the DOGF. From this, with adaptation of
given load balancing policies, we analyzed the average execution times and the deadline violation rates as the execution results
for the client’s request.
... Users should not be network specialists, and moreover the use and the design of their distributed applications must be facilitated. To solve this type of problem, we are interesting in high level tools such as a con guration tool and a dialogue manager 21,26,14,9]. A study for Centaur leads us to de ne a con guration tool based on a library of components provided to the users from which they can build their own application. ...
: In this report we point out the reasons that lead us to distribute modular applications, why we think that modularity is a critical step -- but not a final one -- in the design of distributed applications, and how we distribute. We illustrate our ideas with two experiments developed in the Croap and Secoia projects at INRIA . They concern respectively Centaur and Smeci tools. Key-words: Programming environments, expert systems, distributed systems, communication protocols, client-server architecture. (R'esum'e : tsvp) Anne-Marie.Pinna@essi.fr Wided.Lejouad@sophia.inria.fr Institut National de Recherche en Informatique et Automatique. Unite de recherche INRIA Sophia-Antipolis 2004 route des Lucioles, BP 93, 06902 SOPHIA-ANTIPOLIS Cedex (France) Telephone : (33) 93 65 77 77 -- Telecopie : (33) 93 65 77 65 Le passage d'une architecture modulaire `a une architecture distribu'ee : exemples des syst`emes Centaur et Smeci R'esum'e : Dans ce rapport, nous pr'esentons les raisons ...
The authors describe the performability manager, a distributed system component that contributes to a more effective and efficient use of system components and prevents quality of service (QoS) degradation. The performability manager dynamically reconfigures distributed systems whenever needed, to recover from failures and to permit the system to evolve over time and include new functionality. Large systems require dynamic reconfiguration to support dynamic change without shutting down the complete system. A distributed system monitor is needed to verify QoS. Monitoring a distributed system is difficult because of synchronization problems and minor differences in clock speeds. The authors describe the functionality and the operation of the performability manager (both informally and formally). Throughout the paper they illustrate the approach by an example distributed application: an ANSAware-based number translation service (NTS), from the intelligent networks (IN) area.<
></ETX
ResearchGate has not been able to resolve any references for this publication.