FANTOMAS: Fault Tolerance for Mobile Agents in Clusters

Holger Pals, Stefan Petri, Claus Grewe

Abstract

To achieve an efficient utilization of cluster systems, a proper programming and operating environment is required. In this context, mobile agents are of growing interest as base for distributed and parallel applications. As mobile and autonomous software units, mobile agents can execute tasks given to the system and allocate independently all the needed resources. However, with growing cluster sizes, the probability of a failure of one or more system components and therewith the loss of mobile agents rises. While fault tolerance issues for applications based on "traditional" processes have been extensively studied, current agent environments provide only insufficient, if at all, extensions for a capable reaction on such kinds of failures. We examine fault tolerance with regard to properties and requirements of mobile agents, and find that independent checkpointing with receiver based message logging is appropriate in this context. We derive the FANTOMAS (Fault-Tolerant approach for Mobile Agents) design which offers a user transparent fault tolerance that can be activated on request, according to the needs of the task. A theoretical analysis examines the advantages and drawbacks of FANTOMAS.

Original languageEnglish
Title of host publicationIPDPS 2000: Parallel and Distributed Processing
Number of pages12
Volume1800 LNCS
PublisherSpringer Verlag
Publication date01.12.2000
Pages1236-1247
ISBN (Print)978-3-540-67442-9
ISBN (Electronic)978-3-540-45591-2
DOIs
Publication statusPublished - 01.12.2000
Event15 Workshops Held in Conjunction with the IEEE International Parallel and Distributed Processing Symposium - Cancun, Mexico
Duration: 01.05.200005.05.2000
Conference number: 96623

Fingerprint

Dive into the research topics of 'FANTOMAS: Fault Tolerance for Mobile Agents in Clusters'. Together they form a unique fingerprint.

Cite this