Preemptive Fault Tolerance in DDS based Distributed System using Application Migration
International Journal for Research in Applied Science and Engineering Technology · 2020
Software availability is the main concern of distributed systems regardless of design and framework.Faults are inevitable to any system and system should continue to function even in presence of these faults.Fault tolerance (FT) techniques should make the system fault tolerant by detecting and handling these faults.This paper introduces the technique of proactively anticipating the possible fault and then avoiding that fault by application migration.The possible fault is predicted through sudden or unusual utilization of CPU, RAM, fan speed or CPU temperature.Whenever CPU, RAM, fan speed or temperature breaches the permissible limits of any machine or an alert is generated based on historical data analysis; system takes the necessary steps to redistribute the application load based on software or hardware analytics.The proposed solution has proved to be more efficient than recovery from failure via traditional reactive FT mechanisms such as checkpoint/restart.The use of data distribution service (DDS) as communication middleware has helped in convenient state transfer and state maintenance of migrating application.