Network Redundancy – N10-008 CompTIA Network+ : 3.3

A savvy network administrator includes redundancy options in their network designs. In this video, you’ll learn about active-passive, active-active, diverse paths, and high availability protocols.

One way to implement redundant systems is to purchase two devices but have one of those devices operate at any particular time. If the first device fails, the second device recognizes that the first device is no longer communicating and takes over the process. This usually means that the second device needs to be online, available, and constantly updating its configuration settings so that if that first device fails, it can take over immediately with exactly the same config. Here’s a common active passive configuration for a firewall.

You have an internet provider, there’s one firewall that’s currently active, we have another firewall in the same network configuration that is currently configured as passive. So the first firewall will be passing traffic, the second firewall will be waiting for something to happen. You’ll notice there’s a link between these firewalls, so there’s constant hello messages going back and forth so that both of the firewalls know if the other firewall is available.

Let’s look at some network traffic since this firewall is currently running as the active device. Traffic will pass through normally on its way to the web server. But if something happens to that firewall, let’s say it loses a power supply, that device will no longer be available, and traffic will not be able to pass through that device. The passive firewall recognizes that the first firewall is failed, puts itself into an active mode, and now all network traffic goes through the secondary device to be able to gain access to the web server.

With an active passive configuration, only one of those devices will be operational at any particular time, but you had to buy both devices, so it makes sense that you would want to use both of those devices all the time. This type of configuration is called an active active configuration, because both devices are active simultaneously. However, this also means that the design and implementation of an active active configuration can be a bit more difficult to engineer. For example, we have to consider the data can flow in many different directions.

Some traffic may flow through a device going one direction but through a different device when it’s heading the other direction. This can be a challenge if you need to manage flows, especially if it’s a security device like a firewall. This means you’ll need to monitor and control where the data is going, not only through those active active devices but the other devices that are on either side of those redundant systems.

With an active active configuration, both firewalls are always on and always available. So traffic may pass through one firewall on its way to the web server and other traffic may pass through the other firewall on the way to the web server, and we need to configure these firewalls to recognize that the traffic may flow in any direction at any time, and we need to make sure the configurations are able to be synchronized properly between those devices. In very large data centers, you may have multiple paths to be able to get between point A and point B.

A good example of this is to have multiple internet service providers in and out of the data center. This way, if you lose connectivity to one ISP, you can use the redundant or diverse path to the secondary ISP to maintain your uptime. As you can imagine, this involves more than simply installing a separate network connection. We have to consider the dynamic routing protocols that are in place.

We have to make sure there are failover processes between the two ISPs, and we need to make sure our local devices are configured so that they understand there are two paths to be able to gain access to the internet. But if you spend the time to properly engineer these diverse paths, you’ll now have redundant connections to multiple ISPs, and you’ll be able to failover from one to the other. Of course, diverse paths don’t have to be just to internet service providers. They could be within our own internal network as well.

This is a network that has multiple internet providers, diverse paths into separate firewalls, which also have diverse routers and diverse switches. In this case, we’re connecting to a single load balancer, but we could also engineer multiple load balancers, and ultimately, we have multiple web servers that we connect to. This means we could lose any one of these devices or links and still remain up and running because we have a diverse path that we could follow.

We might also be able to implement redundancy using specialized protocols. Let’s take, for example, the first hop redundancy protocol, or FHRP. If you were to look at the default gateway configuration on your local machine, you’ll notice there is a single IP address that is designated as the default gateway. You may already be thinking that if you lose connectivity to that IP address, then you’ve lost connectivity outside of your local subnet.

But if you’re using FHRP, you can configure multiple routers on your local subnet to all work together. If your default gateway on that subnet was to fail, the first hop redundancy protocol would allow another router to take over that IP configuration and act as the default gateway. Another high availability protocol is the virtual router redundancy protocol, or VRRP.

This is a virtual IP address that’s associated with a router, and if that router fails, that virtual IP address is moved to a different router on that same subnet. This allows you to maintain uptime and availability, even if you happen to change the hardware that’s being used as the router.