Distributed SQL has an administrative overhead compared to a traditional client-server system, which is particularly significant for smaller databases with few queries and transactions. Distributed SQL is therefore primarily suitable for very large databases with numerous transactions, where the overhead is negligible. For very large databases, query times are reduced by the possibility of parallel execution. For fewer than 10,000 queries or 5,000 transactions per second, distributed SQL is usually not the right solution.
Distributed SQL is also less suitable for databases that only contain unstructured data in complex JSON files or fields in JSON format. Although distributed SQL can albania telegram screening also display JSON, the database performance is not very satisfactory for this application area. This is almost certainly an application scenario for a dedicated NoSQL document database.
The same applies to analytical applications: Distributed SQL is not designed for extensive data analysis, but primarily for transactional workloads. Although there are some analytical functions, they are not suitable for complex analytical queries. An analytical database is much better suited for these. In addition, there are all scenarios in which the consistency of the data is not important, such as read-only databases whose data rarely or never changes. In these cases, distributed SQL is probably not suitable as a database solution.