Fast Data Access through Nearest Location-Based Replica Placement

Table 5

Comparison of different techniques.

Mechanism	Main idea	Advantages	Disadvantages
[23]	Three factors are considered as follows: to insert new replicas whose disk space utilization is below average, to insert the new replicas with below-average disk space utilization, to limit the number of recent replica creations on each server, and to spread the replicas across the racks.	High reliability	High energy
		Low response time	High replication cost
		Medium load balancing	High storage cost
		High availability
[24]	The benefits of randomized load balancing are used to improve the data durability.	High reliability	High energy
		Low response time	High replication cost
		High availability	High storage cost
		Medium load balancing
[23]	The trade-off is balanced among different parameters such as mean service time, mean access latency, load variance, energy consumption, and mean file availability. This balancing is done to get near optimal solution.	High availability	High bandwidth consumption
		Low bandwidth consumption
		Low replication cost
[25]	It is the greedy algorithm approach with different start points to find replication node.	High scalability	High energy consumption
		High performance	High replication cost
		Low access latency	High bandwidth consumption
		Low execution time
[26]	This method uses dynamic replica management method, which is based on response time.	High performance	Low reliability
		Low response time	Low load balancing
		High rapid data download	High replication cost
		Low energy consumption
[27]	It is the greedy algorithm-based approach to check whether the application requires higher QoS.	High availability	High time complexity
		High scalability	High bandwidth consumption
		Low replication cost
[28]	It considers cost-effective data replication management as a purpose.	High reliability	High response time
		High availability	Low load balancing
		Low replication cost