Skip to content

amblee0306/tt_assignment_2023

 
 

Repository files navigation

TT_assignment_2023

Tests Test coverage 67.8%

Overview

This GitHub repository contains a completed assignment regarding a backend instant messaging system with one HTTP server and one RPC server. Two APIs are implemented in the system (SEND and PULL). The demo and template for the assignment can be found here. Note that the RPC IDL was not changed.

On top of the template code, the following implementations are completed:

  1. Data storage was added. Redis (Key-Value DB) was utilized.

    • When a message is sent via the send API, the message has to be stored in the database.
    • The key is the chatroom name while the value is the marshalled message.
    • Redis sorted set data structure was utilized (ordered by timestamp of messages).
  2. Pull API was implemented to fetch messages from DB. 

    • Ensured that L messages are returned starting from the cursor, where L refers to the limit passed by the caller. 
    • Ensure that the order of messages returned can be reversed using the reverse parameter in the pull API.
  3. Performance testing.

    • wrk, a modern HTTP benchmarking tool, is utilized.
    • A benchmark for 30 seconds, using 6 threads, and keeping 100 HTTP connections open was run for both send (obtained 1672 requests/sec) and pull (obtained 1243 requests/sec) APIs.

How to use the APIs with sample requests and responses

To send a message to a chat.

  • POST /api/send
  • Response: Nil (unless error)
Param Type Description
chat string Name of chatroom.
text string A text message being sent.
sender string Sender name.
sendtime int64 Timestamp (auto-added when send API is called).

Example curl command:

curl --request POST \
  --url http://localhost:8080/api/send \
  --header 'Content-Type: application/json' \
  --data '{
	"chat": "chatroom12345",
	"text": "hi guys! :) ",
	"sender": "kylie"
}'

To get messages in a chatroom.

  • GET /api/pull
Param Type Description
chat string Name of chatroom.
cursor int64 Starting point (timestamp) of messages to retrieve, 0 by default (if not indicated).
limit int32 The number of messages to return, default 10.
reverse bool To reverse the order of messages returned. If true, the messages will be ordered from oldest to newest.

Example curl command:

curl --request GET \
  --url http://localhost:8080/api/pull \
  --header 'Content-Type: application/json' \
  --data '{
	"chat": "chatroom12345",
	"cursor": 0,
	"limit": 3,
	"reverse": false
}'
  • Response:
Field Type Description
messages array A sorted array of messages.
has_more bool If true, indicates there are more messages in the chatroom that has not been fetch.
next_cursor int64 Starting position (timestamp) of next page if want to fetch more messages. If no more messages, next_cursor will be set to zero.
{
	"messages": [
		{
			"chat": "chatroom12345",
			"text": "hi guys! :) ",
			"sender": "kylie",
			"send_time": 1686481639
		},
		{
			"chat": "chatroom12345",
			"text": "hello to everyone :) ",
			"sender": "jimmy",
			"send_time": 1686481618
		},
		{
			"chat": "chatroom12345",
			"text": "hello from mary :) ",
			"sender": "mary",
			"send_time": 1686481600
		}
	],
	"has_more": true,
	"next_cursor": 1686481600
}

Performance test

The results shown were tested on docker on MacOS with 4 CPU cores and 4GB of Memory.

Pull API

wrk -t6 -c100 -d30s -s ./scripts/wrk-pull.lua http://127.0.0.1:8080

pull_performance

Send API

wrk -t6 -c100 -d30s -s ./scripts/wrk-send.lua http://127.0.0.1:8080

send_performance

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Go 79.2%
  • Thrift 7.8%
  • Shell 3.9%
  • Lua 3.3%
  • Makefile 3.0%
  • Dockerfile 2.8%