villebro commented on issue #2652: URL: https://github.com/apache/datafusion-ballista/issues/2652#issuecomment-6448491206
Here are my thoughts on the topic, though they’re probably heavily influenced by my background with Apache Superset and exploratory analytics/BI in general: For end users, performance and latency are often the most important features of a query engine. If the platform is generally stable and queries return results quickly and reliably, most users probably don’t care much about vcores or how tasks are distributed. Those things are definitely important, but mainly because of how they affect the user experience. I don’t have usage statistics, but I’d wager that pipelined execution could be a good default execution mode. It would mean people trying Ballista get good performance with minimal setup. Obviously we’d want to understand how it behaves under concurrency and failures, and keep blocking execution available for workloads where it’s a better fit. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
