We use cookies to ensure that we give you the best experience on our website. By continuing to browse this repository, you give consent for essential cookies to be used. You can read more about our Privacy and Cookie Policy.

Durham e-Theses
You are in:

Using mobility and exception handling to achieve mobile agents that survive server crash failures

Pears, Simon (2005) Using mobility and exception handling to achieve mobile agents that survive server crash failures. Doctoral thesis, Durham University.



Mobile agent technology, when designed and used effectively, can minimize bandwidth consumption and autonomously provide a snapshot of the current context of a distributed system. Protecting mobile agents from server crashes is a challenging issue, since developers normally have no control over remote servers. Server crash failures can leave replicas, instable storage, unavailable for an unknown time period. Furthermore, few systems have considered the need for using a fault tolerant protocol among a group of collaborating mobile agents. This thesis uses exception handling to protect mobile agents from server crash failures. An exception model is proposed for mobile agents and two exception handler designs are investigated. The first exists at the server that created the mobile agent and uses a timeout mechanism. The second, the mobile shadow scheme, migrates with the mobile agent and operates at the previous server visited by the mobile agent. A case study application has been developed to compare the performance of the two exception handler designs. Performance results demonstrate that although the second design is slower it offers the smaller trip time when handling a server crash. Furthermore, no modification of the server environment is necessary. This thesis shows that the mobile shadow exception handling scheme reduces complexity for a group of mobile agents to survive server crashes. The scheme deploys a replica that monitors the server occupied by the master, at each stage of the itinerary. The replica exists at the previous server visited in the itinerary. Consequently, each group member is a single fault tolerant entity with respect to server crash failures. Other schemes introduce greater complexity and performance overheads since, for each stage of the itinerary, a group of replicas is sent to servers that offer an equivalent service. In addition, future research is established for fault tolerance in groups of collaborating mobile agents.

Item Type:Thesis (Doctoral)
Award:Doctor of Philosophy
Thesis Date:2005
Copyright:Copyright of this thesis is held by the author
Deposited On:08 Sep 2011 18:31

Social bookmarking: del.icio.usConnoteaBibSonomyCiteULikeFacebookTwitter