joaolucasl / nycCams

:camera: QUT 2015 Cloud Computing final assignment

Assignment 2 - Motion detector

Motion server stable version

This version is embeded on AWS/AMI machines. If you update this branch, be sure to update the AMI image by just running misc/Dockerfile on a Ubuntu machine. Remeber to login as lzfelix to download the private lzfelix/cab432 image.

In order to build this project, you are going to need:

• opencv 2.4.x
• node 4.x (on OSX) or 0.12.7 (on Ubuntu)

Notice: nvm doesn't work properly inside Docker containers

Just run npm install on motion_server on the root folder and everything is going to be fine

Tasks

Task	Status
Back end
Implementing motion detection algorithm	Done
Group cameras location on a single j​son​ file	Done
Creating a server which receives a camera request and counts the amount of cars	Done
Require authentication before serving API	Extra / Done
Set up a load balancer to handle load on the processing server	Done
Test manual scaling with this server	Done
Configure automatic scale out/in using the load balancer	Done
Implement database to keep track of most accessed cameras	Done
Remove / route from back server (only before submitting)	To do
: FRONT END :
Collect (lat,long) location of the cameras based on its address	Done
Develop the first page, where the user can select which cameras he wants to monitor	Done
~~Embed Google Maps in this fist page ~~	Done
Retrieve popular cameras from the Database	To do
~~ Embed jQuery on this page to keep statistics updated	Done

DB Login details

mssql -s zw9rn5s0y6.database.windows.net -u nyc_user -p v@c4jairo -d nycdb -e

AWS Service configuration

1. Creating a Virtual Private Cloud (VPC)

Dashboard > VPC > Start VPC Wizard.

Specifications:

• VPC name: nyc_vpc
• Subnet name: nyc_vpcx
• id (current): vpc-f4e55a91
• Gateway id: igw-c823f3

• Route tables: rtb-d8 (custom) / rtb-d930 (main)

2. Create security group

Dashboard > VPC > Security Groups > Create Allows HTTP @ 80 and SSL @ 22 inbound connections

Specifications

• name: nyc_secgroup
• auto assign IP: true

3. EC2 instance details

• network: nyc_vpc
• subnet: nyc_vpcx
• security group: nyc_secgroup
• Use the Docker image in misc/Docker and guestkey here. Don't forget to login as lzfelix
• Modify crontab (crontab -e from this machine)
• Save it as a image. By default aim_nyc_backserver (aim is a typo for AMI)

4. Create load balancer

• Name: nyc_balancer
• Forwarding: port 80 (external) -> port 80 internal (notice Docker container creation)
• Subnet: nyc_vpcx
• Security group: nyc_balancer_secgroup (just allows inbound HTTP connections @ port 80)
• Ping path: '/'

5. Create "Launch Configuration"

EC2 > Auto Scaling > Launch configuration > create

Specifications

• name: aim_nyc_backserver
• AMI: aim_nyc_backserver
• security group: nyc_secgroup

6. Create auto scaling group

• name: nyc_autoscale_group
• configuration: aim_nyc_backserver
• initial size: 2
• max size: 7
• network: nyc_vpc
• subnet: nyc_vpcx
• Health check: 300 (amount of time to consider a instance health to be mapped by the group, in seconds)
• Load balancer: nyc_balancer
• Policies:
  • scale-in: Σ(inbound_data) >= 200.000 bytes for 1 minute in a row
  • scale-out: Σ(inbound_data) <= 50.000 bytes for 1 minute, 3 times in a row