Improved query replica synchronization in Azure Analysis Services is in development
Posted on Tuesday, September 17, 2019
Target availability: Q3 2019
A new Azure Analysis Services setting now in development will improve performance and consistency of query replica synchronization in scale-out environments.
Azure Analysis Services scale-out allows client queries to be distributed across query replicas, reducing response times for high-concurrency workloads.
The new setting will offer these improvements:
- All replicas are synchronized in parallel, reducing the synchronization time considerably.
- All replicas are synchronized in parallel, so data across replicas is much more likely to be consistent during the synchronization process.
- Databases are kept online on all replicas throughout the synchronization process, so users don't need to reconnect.
- The in-memory cache is updated incrementally with only the changed data, which can be much faster than fully rehydrating the model.
Additional memory may be consumed by the query replicas, depending on how much of the in-memory cache needs to be updated. The operation can require up to double the memory because both the old and new segments are kept in memory simultaneously, which is required to keep the database online and available for queries. Replicas have the same memory allocation as the primary query.