An application that performs object detection using an AWS image of a deep learning model.
It can scale automatically based on concurrent requests. The results are stored in S3.
Backend: Java(Spring-boot)
.
Cloud: AWS (EC2, SQS, S3)
.
This application was a course project for CSE-546 (Cloud Computing).
- User hits the static public url of the application (application runs on EC2).
- The app instance puts the request in the queue(SQS) and polls for the response in the output queue.
- The worker app running on other EC2 instances, polls the request in the input queue and performs
deep learning and detects the objects in the video.
- The results are put by the worker in the output queue as well as S3.
- The app instance polls the output queue continuously and when it receives the response, sends it back to
the user.
(The app instance runs on one EC2 instance and starts and terminates the workers based on the number of requests.)