On The Degradation Of Distributed Graph Databases With Eventual Consistency

The 'eventual consistency' approach to updates in a distributed graph database leaves open the possibility that edge information may be corrupted. The process by which this occurs is modeled, with the aim of estimating the time that it takes for the database to become corrupted. A fluid approximation is developed and is evaluated for different parameter settings. Comparisons with simulations show that the results are very accurate.