Pipeline-D

Pipeline-D Setup

• After AWS will grant you access to the private container image, use the command docker run <"AWS Container Image URI"> to launch the container
• Deploy on two or more Linux containers in separate locations: one for the sender, one for the recipient
• Use an encoder on the sender side and a decoder on the recipient side
• For parallel sessions, deploy one container pair per session

Important Notes

• Runs in a closed and isolated cloud environment
• Handles one request at a time
• Assumes input is valid
• Invalid data may cause data leakage, corruption, denial of service, or remote code execution

Usage Overview

• Listens on HTTP port 8080
• Maintains a stateful session per sender and recipient
• Each session builds a dictionary in memory to improve performance
• Reset sessions to release memory when needed

Message Flow

• Sender resets encoder if needed
• Sender sends message to encoder
• Encoded message is sent to recipient
• Recipient resets decoder if sender did
• Recipient sends encoded message to decoder
• Decoder sends the result to its final destination
• Repeat steps two through five for the entire session

API Reference

• POST compress

• Compresses JSON using internal methods

• Input

• Content-Type is application slash json

• Body must be valid JSON

• Output

• Content-Type is application slash octet-stream

• Body is compressed data

• Ensure JSON is valid

• No trailing commas

• Strings must be properly escaped

• Do not include NaN or Infinity values

• POST decompress

• Restores compressed data to original JSON

• Input

• Content-Type is application slash octet-stream

• Body is encoded data

• Output

• Content-Type is application slash json

• Body is the restored JSON

• Decode messages in the same order as they were encoded

• Do not mix data from different sessions

• POST reset

• GET reset

• Resets the session and clears memory

• No input required

• Returns HTTP 200 OK

• Can be used to confirm service is ready