I will use Sparsity (former name is DEX) graph database in my case. But haven't got experience related with graph databases. I have read several articles, many of them suggest several things(Rexster server, REST API, Bluprint etc.) in order to use those DB-s. So, is it necessary to use Server(i.e Rexster) for those DB-s? And what Configurations I must do in order to do DB operations. I will use Java as a programming language.
The primary purpose of Rexster is to provide a way for non-JVM languages to communicate with a Blueprints Graph. I would say that it is also useful for centralizing access to graph traversals as a service with or without the JVM in mind. It is not explicitly required to use the TinkerPop stack or any specific graph database implementation of Blueprints. If you are building a JVM application you may choose to simply reference TinkerPop libraries and utilize them directly as dependencies.