Amazon Elastic Load Balancing is used to automatically scale up and down the number of server nodes we have. For example, if we wanted to provision an EC2 machine as our website server, this one machine might not be able to handle an increase in traffic.
If the traffic increases, AWS will automatically add one more machine, which is an exact copy of the other machine, and that will take the additional node. With Amazon Elastic Load Balancing, you set the minimum and maximum limits. For example, at minimum, you should have two servers; at maximum, you can have four servers. So the number of servers can oscillate between two and four depending on the traffic.
The most valuable feature of Amazon Elastic Load Balancing is scaling. We just have to enable the load balancer, set the minimal and maximum limits, and it will handle everything by itself.
We faced a problem once. The different machines created by Amazon Elastic Load Balancing have separate IP addresses. These machines, in turn, connect to our MongoDB Service. Since MongoDB has an IP address-based firewall in the middle, it will only allow IP addresses you have whitelisted.
So whenever the solution creates a new machine, we have to add the IP address to our MongoDB whitelist. Whenever it deletes and recreates, the old IP address becomes invalid. This causes our connections to break, and our customers don't get the data. The machines created by Amazon Elastic Load Balancing have different IP addresses, which we are not able to whitelist or predict.
My organization has been using Amazon Elastic Load Balancing for more than one year.
Two people are using the solution in our organization.
Amazon Elastic Load Balancing’s initial setup is very easy.
We use the solution on-cloud. A load balancer's auto-scaling service is always needed when we set up a dedicated server because we cannot do the scaling manually. We used a load balance for our server-based approaches, but we don't need a load balancer for our serverless approaches.
Overall, I rate Amazon Elastic Load Balancing nine out of ten.