Group members:
- Deval Pandya - 1225424200
- Karthik Ravi Kumar - 1225910467
- Tirth Hihoriya - 1225413475
Accound id: 477824770261
IAM user name: Faculty
Password: ccprojectfaculty#1
Web-tier's URL: http://3.92.213.222:8080/upload
SQS queue names: RequestQueue and ResponseQueue
S3 bucket names: cc-project-input and cc-project-output
- I have designed the end to end flow of this project, like determining the AWS components to be used and setting them up like SQS Queues, S3 Buckets, Security Groups, EC2 instances and their user data, IAM Instance profiles. I also played a part in designing the logic for Web Tier, App Tier, and Controller. I implemented the code for the controller which is used for Autoscaling of App Tier instances as per the number of messages in the Request Queue. I did the testing for the whole application which involved sending several requests (Single/Concurrent) from the workload generators provided to us and optimized the algorithms of App Tier and Web Tier in order to showcase the features of autoscaling in our project.
- I have designed the Web tier. The listener part of the web tier includes receiving messages from the user, storing it and encoding the images. Other parts include setting up and sending messages to the request queue,testing the request queue, receiving messages from the response queue and printing the output.
- I have designed the App-Tier. The app-tier includes basic image classification program. It receives a task from the RequestQueue and processes it. Once it is done with processing, it sends the output to the Response Queue. It also stores the input in input-bucket and output in output-bucket. If no new task comes to it, it waits for 5 seconds. I have tested the app-tier code rigorously and updated it to have more robust code. In the code there are many points where there is a chance of exceptions. These exceptions may occur due to reasons such as, the code might not get a message from RequestQueue, the input and output might not be stored in S3 buckets successfully, the message form RequestQueue might not be deleted successfully after it gets processed, etc. Thus, to handle it the code is taken care of with the ‘try-except’ block. If any exception occurs then it will wait for 5 seconds and will try to process again.
Login with your ASU id to see it.