Elastic Load Balancing & Auto Scaling Groups

• Scalability: Ability to accommodate a larger load by making hardware stronger (scale up) or by adding nodes (scale out)

  • Vertical scalability: Implies non-distributed systems, such as a database
    • Increasing size of the instance
  • Horizontal scalability: Implies distributed systems
    • Increasing the number of instances / systems for your application

• High Availability: Running your application / system in at least 2 AZ

  • Usually goes hand-in-hand with horizontal scaling
  • Purpose is to mitigate damage from disasters, such as a data centre loss

• Elasticity: Given a scalable system, elasticity implies there's auto-scaling such that the system scales based on the load

  • Pay per use, match demand, optimised costs

• Agility: New IT resources are easily attainable, reducing the time needed to make resources available to developers from weeks to minutes

Elastic Load Balancing

• Load balancers: Servers that forward internet traffic to multiple EC2 instances downstream

• Benefits of using a load balancer:

  • Spread load across multiple downstream instances
  • Expose a single point of access (DNS) to your application
  • Easily handle failures of downstream instances
  • Do regular health checks to your instances
  • Provide SSL termination (HTTPS) for your websites
  • High availability across zones

• Elastic Load Balancer (ELB): A managed load balancer

  • AWS guarantees it will be working, takes care of upgrades, maintenance and high availability
  • But AWS only provides a few configuration knobs

• Costs less to set up a load balancer but more effort (maintenance, integrations)

• AWS offers 4 kinds of load balancers:

  • Application Load Balancer (HTTP / HTTPS / gRPC): Layer 7
    • Useful scenarios: Need HTTP Routing features, for static DNS (need static URL)
  • Network Load Balancer (TCP / UDP / TLS): Layer 4
    • High performance, can handle millions of requests per second with low-latency
    • Useful scenarios: Static IP through Elastic IP
  • Gateway Load Balancer (GENEVE on IP packets): Layer 3
    • Useful scenarios: Route traffic to firewalls that you manage on EC2 instances (classic firewall, intrusion detection, deep packet inspection)
  • Classic Load Balancer (retired in 2023): Layer 4 & 7

Auto Scaling Group

• Load on websites and applications can change, the cloud enables you to create and remove servers quickly

• Goal of Auto Scaling Group (ASG):

  • Scale out (add EC2 instances) to match an increased load
  • Scale in (remove EC2 instances) to match a decreased load
  • Ensure we have a minimum and maximum number of machines running
  • Automatically register new instances to a load balancer
  • Replace unhealthy instances

• Only runs at an optimal capacity which saves costs

• Scaling strategies are useful when your load has predictable time-based patterns

  • Manual Scaling: Update the size of an ASG manually
  • Dynamic Scaling: Respond to changing demand
    • Simple / Step Scaling:
      • Ex. Given a minimum and maximum threshold representing load, when exceeded, either add or remove servers
    • Target Tracking Scaling:
      • Ex. Want the average ASG CPU to stay at around 40%
    • Scheduled Scaling: Anticipate a scaling based on known usage patterns
      • Ex. Increase minimum capacity to 10 at Friday 5pm
  • Predictive Scaling: Using machine learning to predict future traffic
    • Automatically provisions the correct number of EC2 instances in advance