A Survivable Distributed Database Against Byzantine Failure

Abstract

Distributed Database Systems have been very useful technologies in making a wide range of information available to users across the World. However, there are now growing security concerns, arising from the use of distributed systems, particularly the ones attached to critical systems. More than ever before, data in distributed databases are more susceptible to attacks, failures or accidents owing to advanced knowledge explosions in network and database technologies. The imperfection of the existing security mechanisms coupled with the heightened and growing concerns for intrusion, attack, compromise or even failure owing to Byzantine failure are also contributing factors. The importance of survivable distributed databases in the face of byzantine failure, to other emerging technologies is the motivation for this research. Furthermore, It has been observed that most of the existing works on distributed database only dwelled on maintaining data integrity and availability in the face of attack. There exist few on availability or survibability of distributed databases owing to internal factors such as internal sabotage or storage defects. In this paper, an architecture for entrenching survivability of Distributed Databases occasioned by Byzantine failures is proposed. The proposed architecture concept is based on re-creating data on failing database server based on a set threshold value.The proposed architecture is tested and found to be capable of improving probability of survivability in distributed database where it is implemented to 99.6% from 99.2%.