Long description

Year 2006:

User uploads image using upload button on H T M L form.

The image file is posted to server handler using H T T P.

The image file is then saved in a folder on the server using about 12 lines of P H P.

Year 2013:

User uploads image using upload button on H T M L form.

JavaScript provides content checking and preview.

The image file is posted to server handler using H T T P.

P H P handler verifies user token by checking against database.

The image file from the server is saved within database and associated with user record.

Year 2020:

User drags and drops image onto H T M L canvas.

JavaScript handles the user event and provides preview.

The image file is posted to server handler using H T T P.

The server handler verifies user's token using Authentication A P I.

Node handler saves the file in cloud file store open parenthesis such as “A” W S S 3 close parenthesis.

Cloud listener is triggered when a new file is added to file store.

The cloud listener triggers the execution of several cloud functions within job queue or message pipeline open parenthesis such as Rabbit M Q or “A” W S S Q S close parenthesis.

The images are compressed and thumbnails are generated for different device sizes open parenthesis C plus plus close parenthesis.

The thumbnails are stored within cloud based no - S Q L database open parenthesis such as Dynamo D B or FireBase close parenthesis.

Image recognition A P I s are invoked to categorize image contents open parenthesis Python close parenthesis.

The results are also stored in cloud D B.

The thumbnails are sent to C D N open parenthesis Content Delivery Network close parenthesis for improved latency.

Push notifications are sent to subscribers of user’s image feed open parenthesis using WebSockets and push service such as PubNub close parenthesis.

Finally, a record of each stage is saved in transaction log which later will be analyzed using MapReduce based tools like Hadoop.