Although graph databases have extensively found applications in the relationship-centered era, a time-version support is seldom provided. While current storage systems capture the most recently updated snapshot of the underlying graph, most real world graphs embed a dynamic behavior translating the fact that vertices or edges can join or leave the graph at any time instant. Regarding that, a graph database should faithfully maintain the state of every graph's element permitting the analysis and prediction of the underlying system's performance. Since physical deletions are forbidden in such a scenario, the outgrowing size of data is a crippling restriction steering the interest in this area towards the optimization of the persistent storage. However, capturing and storing the state of the graph as full snapshots adds a storage overhead traded by faster query responses. Accordingly, the choice of an appropriate storage engine should be adapted with the threshold of accepted query latencies and the available storage resources. This paper will recognize the anterior academic work in the era of temporal graph databases while highlighting the existing tradeoff between storage and computation time costs. The implementation of GDBAlive, a temporal graph database using two state-of-the-art techniques Copy+Log and Log, is provided relying on a robust column oriented data store. In order to optimize the responsiveness of temporal queries in terms of computation times, we will introduce two fetching strategies "AsyncFS" and "Forced Fetch" and prove their efficiency on a real dataset.