Jellyfish-130 App Optimization System Design

Optimization Overview

• Identified system bottlenecks, modified system configurations, added an NGINX load balancer, cached the data, indexed relational keys, and horizontally scaled a heavily READ / GET backend to handle network requests and reduce latency for a 60 million records PostgreSQL database.
  • Resulted in:
    • Increased read throughput by 57%
    • Reduced response times to an avarege of ~83ms for image retrieval operations while maintaining a 0% error rate for 1,750 requests per second, or ~105,000 requests per minute
      • Stress test data in Loader.io: https://bit.ly/2HO2Szs

Table of Contents

• Client Demo
• API Endpoints Documentations
• Installing Dependencies
• Getting Started
• Getting Deployed
• Database Schema Design

Demo

Technologies / Frameworks Used

• PostgreSQL
• Cassandra
• Express.js
• React
• Node.js
• Styled Components

Other Tools: NGINIX, EC2, S3, Loader.io, New Relic, K6

API-CRUD-Operations

For Guests / Site Visitors:

Get all listings on the platform upon ComponentDidMount()

• READ / GET: /listings
• Path Parameter(s): none

Request Body:

  None

Response object:

{
  [
    {
    listingID: number,
    title: string,
    location: string,
    description: string,
    quote: string,
    hostname: string,
    rating: string,
    guests: number,
    bedrooms: number,
    beds: number,
    bathrooms: number,
    superhost: boolean,
    reviews: array,
    gallery: object {
        featured: array [],
        rooms: [
          {
            name: string,
            amenities: array [],
            images: array [
              object {
                url: string,
                comment: string
              }
            ]
          }
        ]
      }
    },
    {
      ... second listing object
    },
    {
      ... third listing object
    },
    {
      ... so on
    },
  ]
}

Sucess Status Code: 201

For Hosts / Listers:

Add a listing to listings

• Create / Post: /listings
• Path Parameter(s): none

Request Body:

{
      listingID: number,
      title: string,
      location: string,
      description: string,
      quote: string,
      hostname: string,
      rating: string,
      guests: number,
      bedrooms: number,
      beds: number,
      bathrooms: number,
      superhost: boolean
}

Response Object:

{
  listingId: number
}

Success Status Code: 201

Add photo(s) to a listing

• CREATE / POST: /listings/:listingId/photos/
• Path Parameter(s): :listingId

Request Body:

{
  images: array []
}

Response Object:

  Success Status Code: 200

Get a listing on the platform

• READ / GET: /listings/:listingId
• Path Parameter(s): none

{
  listingID: number,
  title: string,
  location: string,
  description: string,
  quote: string,
  hostname: string,
  rating: string,
  guests: number,
  bedrooms: number,
  beds: number,
  bathrooms: number,
  superhost: boolean,
  reviews: array [],
  gallery: object {
      featured: [],
      rooms: [
        {
          name: string,
          amenities: array [],
          images: array [
            object {
              url: string,
              comment: string
            }
          ]
        }
      ]
    }
}

Response Object:

  Success Status Code: 200

Update photo(s) to photos for a listing

• UPDATE / PUT: /listings/:listingId/photos/
• Path Parameter(s): :listingId

Request Body:

{
  url: string
}

Response Object:

  Success Status Code: 201

Delete a photo from photos from a listing

• DELETE / DELETE /listings/:listingId/photos/:photoId
• Path Parameter(s): :listingId, :photoId

Request Body:

  None

Request Object:

  Sucess Status Code: 204

Dependencies

From within the root directory:

npm install

Requirements

• Node 10.5.0
• npm 6.14.7
• MongoDB 4.2.8

Development

From within the root directory:

To run dev environment/webpack

npm run build

To run server

npm start

To seed database

npm run seed

To run tests

npm run test

Usage

To get started with the Gallery Module, run http://localhost:3001/ on your local web browser.

Database Schema Design

PostgreSQL:

CREATE SCHEMA gallery_service;

CREATE TABLE gallery_service.listings (
  listing_id SERIAL PRIMARY KEY,
  title varchar(55),
  listing_name varchar(55),
  rating varchar(4),
  num_reviews smallint not null,
  superhost boolean,
  listing_location varchar(55),
  hostname varchar(55),
  host_avatar_url text,
  share boolean,
  save_feature boolean
);

CREATE TABLE gallery_service.feature_photos (
  feature_photo_id SERIAL PRIMARY KEY,
  listing_id integer not null references gallery_service.listings(listing_id),
  photo_1_url text,
  photo_2_url text
);

CREATE TABLE gallery_service.rooms (
  room_id SERIAL PRIMARY KEY,
  listing_id integer not null references gallery_service.listings(listing_id),
  room_name varchar(55)
);

CREATE TABLE gallery_service.photos (
  photo_id SERIAL PRIMARY KEY,
  listing_id integer not null references gallery_service.listings(listing_id),
  room_id integer not null references gallery_service.rooms(room_id),
  photo_url text,
  photo_caption varchar(500)
);

Cassandra:

CREATE KEYSPACE gallery_service WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 1 } AND durable_writes = 'true';

USE gallery_service;

CREATE TABLE listings_by_id (
  listing_id int,
  listing_title text,
  listing_name text,
  listing_rating decimal,
  num_reviews int,
  superhost boolean,
  listing_location text,
  hostname text,
  host_avatar_url text,
  share boolean,
  save_feature boolean,
  PRIMARY KEY(listing_id)
);

CREATE TABLE feature_photos_by_listings_id (
  listing_id int,
  feature_photo_id int,
  photo_1_url text,
  photo_2_url text,
  PRIMARY KEY(listing_id, feature_photo_id)
);

CREATE TABLE photos_by_listing_id (
  listing_id int,
  room_id int,
  photo_id int,
  room_name text,
  photo_url text,
  photo_caption text,
  PRIMARY KEY(listing_id, room_id)
);