Purpose:
- Optimize the Elasticsearch for IoT applications (huge volume of streaming data).
- search in specific section not whole of data store.
Please look at the GridDB architecture (a key-container No-SQL database) for the IoT scenarios.
I think it would be great if the elasticsearch can also store data as key-container (key=thing UUID, container = streaming data of specific thing in ONE document not). I think this architecture improves the search performance even more than now. because you only need ONE document for each IoT device (consider it as an additional block) and can enhance search/usage applications.
Example:
In IoT scenarios, often we need to query and explore in a specified category not whole of data store. for example, some customers have only specific range of IoT devices, so they should be able to search for only that category and its not necessary to search in whole of data store which is very huge.
Related issue: https://github.com/elastic/elasticsearch/issues/55841
What do you think about it?