Elastic Load Balancing & Auto Scaling Groups

Question 1

Scalability

Answer 1

• App/system can handle greater loads
• Accommodate larger load by making the hardware stronger (scale up), or adding nodes (scale out)

Question 2

Vertical Scalability

Answer 2

Increase the size of the instance

Question 3

Horizontal Scalability

Answer 3

Increase number of instances / systems

Question 4

High Availability

Answer 4

• Associated with Horizontal Scalability
• Run in at least two AZs
• Survive a data center loss/disaster

Question 5

Elasticity

Answer 5

• Sone “auto-scaling” so the system can scale based on the load
• “cloud-friendly”
• pay-per-use
• match demand
• Optimize costs

Question 6

Agility

Answer 6

• Not related to scalability
• New resources are only a click away
• Reduce time to make those resources available from weeks to minutes

Question 7

Load Balancing

Answer 7

Server forwards internet traffic to multiple servers (EC2 Instances) down stream

Question 8

Why use a Load Balancer

Answer 8

• Spread load
• Expose a single point of access
• Seamlessly handle failures
• Do regular health checks
• Use across AZ

Question 9

Elastic Load Balancer use case

Answer 9

• Managed service AWS will:
  Guarantee it will be working
  Take care of upgrades, maintenance, high availability
  Provide a few configuration knobs
• Cost less to setup than your own, and with much less effort.

Question 10

Types of Elastic Load Balancers

Answer 10

• Application -Layer 7
• Network (ultra-high performance, TCP/UDP) - Layer 4
• Gateway - Layer 3
• Classic (retired 2023) - Layer 4 & 7

Question 11

Application Load Balancer

Answer 11

• HTTP / HTTPS
• gRPC
• Layer 7
• HTTP routing features
• Static DNS (URL)

Question 12

Network Load Balancer

Answer 12

• TCP / UDP
• High performance
• Static IP with Elastic IP

Question 13

Gateway Load Balancer

Answer 13

• GENEVE Protocol on IP Packets
• Layer 3
• Route traffic to firewalls managed on EC2 Instances
• Routes to 3rd party security virtual appliances

Question 14

Auto Scaling Groups

Answer 14

• Add/Rem EC2 instances with volume
• Have a min/max # of machines
• Auto reg/dereg instances to load balance
• Replace unhealthy instances
• Cost Saving, run at optimal capacity

Question 15

Manual Scaling

Answer 15

Update size of ASG manually

Question 16

Dynamic Scaling

Answer 16

• Respond to changing demand
• Type
  Simple/Step Scaling
  Target Tracking Scaling

Question 17

Simple/Step Scaling

Answer 17

Whenever a CloudWatch alarm is triggered

Question 18

Target Tracking Scaling

Answer 18

• Based on target performance
• Like keeping CPU usage at 40%

Question 19

Schedule Scaling

Answer 19

*Anticipate scaling based on known usage patterns

Question 20

Predictive Scaling

Answer 20

• Machine Learning to predict future traffic (algorithm)
• Automatically provision the right # of EC2 instances in advance
• Useful for predictable time-based patterns