Low latency databases with high query volumes and large numbers of connections often have multiple redundant communication paths from clients to database servers. When queries fail, clients that submit queries have difficulty identifying the source of query failures and are unable to take actions to ensure high performance and service availability. This disclosure describes database techniques to tune performance and identify reasons for failed queries. A metadata layer within the database is used by the database server to provide anticipated time-to-serve based on the current server load and the complexity of the query. Clients can use the time-to-serve data to plan queries, set dynamic timeouts, detect network black holes, attempt alternate communication pathways, etc. Further, clients can share amongst each other the current time-to-serve and server health, enabling multiple clients to plan respective queries based on the response by the server to a single client. Similar techniques apply to identify communication failures generally in data transfer situations.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Gogel, Ed, "Improving Speed and Reliability of Database Queries", Technical Disclosure Commons, (January 03, 2023)