In this video, Sharon outlines the different Azure load balancers; Internal and Internet. Probes, load balancing rules and NAT will also be covered.
- [Instructor] Azure Load Balancers create a highly available structure by distributing traffic among healthy virtual machines within the network. In Azure, we have two load balancers. We have an Internet Load Balancer and an Internal Load Balancer. In our example here, I'm showing you both load balancers. As we work from the top down, traffic would enter our first load balancer, which is our Internet Load Balancer. This would have a public IP address associated with it. The traffic would then be distributed across the virtual machines within an availability set.
Traffic that then needed to be directed to our lower database tier would pass through an Internal Load Balancer and the same process would be repeated. The traffic would be distributed equally among the virtual machines within that tier. These virtual machines are in an availability set. One thing you will notice here, though, with our Internal Load Balancer, we have another connection. This would be for our external users, let's say our on-premise environment. These users would connect directly into the Internal Load Balancer and then their traffic would be distributed across those virtual machines, instead of them going out through their gateway and then coming back in through our Internet Load Balancer.
The load balancers will only distribute traffic to healthy virtual machines, or instances, within the balance set. We use Probes to do this, and they just monitor those virtual machines. If a machine is deemed unhealthy, new requests will not be sent to it, and the traffic will continue to be directed to the healthy virtual machines. Machines are deemed unhealthy if HTTP 200 OK is not returned within the specified timeout. The endpoints of virtual machines are probed every 15 seconds and the default timeout is 31 seconds.
If the load balancer does not receive the response within that timeout, it will be deemed unhealthy. And we have two types of probes. We have an HTTP Probe and a TCP Probe. When you use an HTTP Probe, the instance, or virtual machine will be deemed unhealthy if the application responds with a code that is not 200, or if a response hasn't been received within the specified timeout or if TCP is reset. A TCP Probe uses a three-way handshake to initiate communication. And just like the HTTP Probe, if the server does not respond in the specified timeout period, or TCP is reset, the virtual machine or instance will be deemed unhealthy.
Within our load balancers, we also have load balancing rules. Basically, this directs traffic based on a port. We can also enable session persistence within the load balancing rules. This guarantees that the client will always connect to the same virtual machine instance during the session. We can also enable floating IP, or sometimes you'll hear it referred to as Direct Server Return. This would be enabled when configuring a SQL Always On availability group listener. Otherwise, leave the default of disabled. And finally, we can configure NAT within our load balancers.
NAT will allow us to direct specific traffic to a specific VM. For example, you have clients on the internet who need to RDP into that virtual machine, you could set the rule up in the load balancer that when that traffic on that port enters the load balancer, it is then directed to the corresponding port on the VM. And that's the overview for load balancers within Azure.
- Creating an Azure virtual network
- Creating a virtual network using PowerShell
- Deploying a VM into a virtual network
- Modifying IP addresses
- Working with Azure DNS
- Configuring NSGs
- Setting up load balancers
- Configuring Azure load balancers
- Creating an application gateway
- Setting up on-premises connectivity
- Adding gateway VPNs
- Validating VPN devices
- Configuring VNet
- Creating site connections