Cloud Computing

 AWS Load Balancers: Load balancer is a service that uniformly distributes the network traffic and workloads across multiple servers or a cluster of servers. Load balancer increases the availability and fault tolerance of an application. 

Elastic Load Balancer (ELB) is a load balancing service for the AWS deployments and scaling itself as necessary to handle the load. Incoming traffic is distributed across EC2 instances in multiple availability zones and ELB becomes the single point of contact for clients.

Creating Application Load Balancer:

• Go to AWS Management Console and click on the Services drop-down
• choose EC2 and Scroll down to select ‘Load Balancers’
• Select Create button under Application Load Balancer
• Configure Name, create a new VPC, add a tag; select mapping, default subnet, use weighted routing in the listener rules and then choose launch

Autoscaling: Scaling is the process of adding/removing capacity/resources as needed. Vertical and Horizontal are two distinct scaling types. Scale out is adding the capacity/resources and Scale in is removing the capacity/resources. Autoscaling can dynamically increase or decrease capacity as needed.

 Autoscaling group contains a collection of EC2 instances that identical. autoscaling group needs  launch configuration to be specified which cannot be changed. All new instances launched will use configuration and scaling policies will determine EC2 instances launching and termination. 

Launch configuration is a template that is specified with Autoscaling groups to launch EC2 instances. It can be created by either a) From scratch: Image ID, instance type, storage devices, etc. b) From an EC2 instance: Attributes from the instance are copied. Block device mapping of the AMI is included.

Scaling policies and alarms mention how to scale, and alarms decide when to scale. Amazon CloudWatch enables policies and monitor metrics e.g., CPU utilization, etc. When the threshold is breached, scaling policies are executed with pre-defined min, max  and the desired capacity. Scaling can also be achieved based on a schedule i.e scale at a given time and date. Finally, Scaling can be done based on demand i.e. scaling occurs when the CPU utilization of the current running instances grows beyond a fixed usage limit.