Fragmentation, allocation and replication are techniques widely used in relational databases to improve the performance of operations and reduce their cost in distributed environments. This article shows an analysis of different methods for database fragmentation, allocation and replication and a Web application called FRAGMENT that adopts the work technique that was selected in the analysis stage, because it presents a fragmentation and replication method, it is applied to a cloud environment, it is easy to implement, it focuses on improving the performance of the operations executed on the database, it shows everything necessary for its implementation and is based on a cost model. FRAGMENT analyzes the operations performed in any table of a database, proposes fragmentation schemes based on the most expensive attributes and allocates and replicates a scheme chosen by the user in a distributed environment in the cloud. This work shows a common problem in fragmentation methods, overlapping fragments, and provides an algorithm with an approach to address it. This algorithm results in the predicates that will define each fragment in a distributed environment. To validate the implemented technique, a second web application is presented, dedicated to simulate operations on sites and focused on producing a log file for the main application. Experiments with the TPC-E benchmark demonstrated lower response time of the queries executed against the distributed database generated by FRAGMENT compared with a centralized database.