| Summary: | In distributed systems such as grid and cloud, effective data management has become a
critical requirement. In achieving effective data management, the data replication
technique is a common solution to address metrics such as availability, bandwidth
consumption, resource consumption, response time, and storage usage. However,
reducing response time and resource consumption like computation and cloud storage
became major challenges within the cloud environment for data replication. Effective
data management may be improved by reducing the time consumption and reducing
resources in searching data replicas location in a cloud environment. Consequently, the
performance can be increased, and therefore, the cost of updating a new replica may be
reduced. Choosing data replica locations on the unknown dynamic cloud environment
is hardness or NP-hardness problem (Non-deterministic Polynomial-time). For this
reason, the search for a new meta-heuristic algorithm is still a useful endeavour. This
thesis presents the design and implementation of a new meta-heuristic algorithm based
on Markov Chain Monte Carlo MCMC and BAT Algorithm, called Monte Carlo
BAT Optimization (MCBO), for generating a random start location for the replication
route and determining the most effective path.MCBO is the first approach that uses
BAT as the backbone engine to introduce Markov Chain Monte Carlo replication
techniques and to optimize the dynamic data replication on a cloud environment for the
reduction of time and resources in unknown areas. The experimental results supported
by non-parametric statistical analysis demonstrate that the MCBO algorithm gives
competitive performance over its counterparts, MCBO has achieved 58 % of time
reduction, which indicates reduce time and resource consumption to distribute replica
on a cloud environment, benchmark with results 48% for GA (Genetic Algorithm) and
46% PSO (Particle Swarm Optimisation). This fmding contributes to the optimization
data replication technique by minimizing the time and resource consumption in the
cloud environment. Till now, there is no prior work that incorporates MCMC and BAT
algorithms to minimize time and resource consumption in the cloud environment. This
research contributes to the field of dynamic replication strategies as it provides a new
optimization model called MCBO. MCBO enhanced data replica distribution by
reducing time and resource consumptions and providing the most optimum path to
update data replica. In the cloud computing environment, the experimental results
indicate a considerable improvement in replication strategy.
|
|---|