An efficient causal ordering algorithm for multicast communication channels

In this paper, we present an efficient causal algorithm that can be used in multicast communication environments, in particular the overlapping multi-channel case, where a participant can belong to and communicate through more than one multicast channel. The causal algorithm is built on the paradigm...

Descripción completa

Detalles Bibliográficos
Autores: SAUL EDUARDO POMARES HERNANDEZ, LUIS ALBERTO MORALES ROSALES
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2010
País:México
Institución:Instituto Nacional de Astrofísica, Óptica y Electrónica
Repositorio:Repositorio Institucional del INAOE
Idioma:inglés
OAI Identifier:oai:inaoe.repositorioinstitucional.mx:1009/1413
Acceso en línea:http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/1413
Access Level:acceso abierto
Palabra clave:info:eu-repo/classification/Immediate dependency relation/Immediate dependency relation
info:eu-repo/classification/Causal ordering/Causal ordering
info:eu-repo/classification/Group communication/Group communication
info:eu-repo/classification/Distributed systems/Distributed systems
info:eu-repo/classification/cti/1
info:eu-repo/classification/cti/12
info:eu-repo/classification/cti/1203
Descripción
Sumario:In this paper, we present an efficient causal algorithm that can be used in multicast communication environments, in particular the overlapping multi-channel case, where a participant can belong to and communicate through more than one multicast channel. The causal algorithm is built on the paradigm of group communication. The groups are established according to the participant channel subscription. In order to reduce the amount of control information (CI), we propose an extension of the immediate dependency relation (IDR), which was introduced by Peterson in the context of one group. This IDR extension allows us to define necessary and sufficient CI to ensure causal delivery in a multi-group environment. We show that through the use of the IDR extension, we reduce the amount of CI sent per message without imposing restrictions on interaction or execution (e.g. network topology, rediffusion servers, execution models, etc.). These characteristics allow our algorithm to be suitable for use in large distributed decentralised systems. We show the efficiency of our causal algorithm in terms of the overhead timestamped per message.