Many of AWS services and innovations are customer-driven. The AWS Gateway Load Balancer service is no exception. Many customers have relied on virtual appliances from AWS partners and the AWS marketplace. However, the deployment process and scaling for virtual appliances was difficult to say the least. First, we will need to be able to direct all traffic, inbound and outbound, from an Internet Gateway or Virtual Private Gateway to an Elastic Network Interface of a specific EC2 instance in a VPC. 

This feature is essential and happens to be implemented using VPC Ingress routing for the Internet Gateway. Using a VPC Ingress routing, we can forward traffic to a Gateway Load Balancer by updating the route tables in a VPC. The next feature needed is to deal with IP tunneling, such as not to incur errors or conflicts with IP addressing. In a nutshell, we need to be able to grab all traffic, inbound and outbound for the VPC, and redirect it to a virtual network appliance for security processing, and not interrupt the normal flow and interactions of the request and the response. 

The process needs to be transparent. The AWS Gateway Load Balancer uses a single point of access for all inbound and outbound traffic, and allows you to scale your virtual appliance with demand as done with other Elastic Load Balancers, like the Application Load Balancer. The Gateway Load Balancer routes traffic through healthy virtual appliances and stop sending traffic if an appliance becomes some healthy. 

Using Gateway Load Balancer, you can also add your own logic into any networking path in AWS when you want to inspect and take action on packets. The AWS Gateway Load Balancer sends inbound and outbound traffic transparently over the same consistent route and using the same target. This implements sticky, transparent, and symmetric flow.