In the current version of the API, we have implemented a very simple and a naïve way to handle those requests. We have:
- Implemented a queue (acting like a message queue) to store all those requests
- Batch every 5-6 concurrent requests and then treat as one request to put into
fastdtw algorithm.
- And then return the response to the frontend.
But we need some better techniques to do that more effectively and in an optimized manner. Some ways to get started with are:
- Getting started with
Apache FLink to handle those requests.
- Implementing Async message queues
etc