A Comparison of Relational and Graph Databases for CRM Systems

In an age of increasing usage of heavily interconnected platforms, graph databases have increased in popularity due to their ease of modelling these systems. This thesis will evaluate if the graph database Neo4j can be used to model Lime's CRM System efficiently by comparing the performance of Neo4j and MS SQL on queries similar to those existing in Lime CRM. To benchmark the databases, they were created with the core entities of the Lime CRM system. The databases are created with the same structure and schema as in Lime CRM. Thereafter, the databases were populated with randomly generated data. These databases were then queried a large amount of times with random values in every query to prevent caching. Our results show that MS SQL was 4-10 times faster for insertion queries. Furthermore, MS SQL was also 2-5 times more efficient at handling queries that return many rows/nodes of data.Det finns flertal fall då grafdatabaser överträffar traditionella relationsdatabasers pre-standa. Är grafdatabaser potentiellt den nya lagringsmetoden vid stor mängd sam-mankopplad data? Vilka villkor krävs? För att sätta det på prov har vi skapat entestmiljö med en stor mängd relationsbunden data