Load balancing and clustering can be used to prevent downtime caused by patching systems. For example, if you have a mission-critical database server that must never go down, then create a cluster for it. Then when you need to patch the server or its database application, you can fail over each node in the cluster and patch each node separately while it's offline.
Or say you have a web server that must never go down, even for a reboot due to a patch being applied. For this scenario you could create a load balanced web farm, then take each web server offline and patch them individually and return them to the farm.
Sure these scenarios are more complex to implement and expensive too, but if your business demands zero downtime then you should implement something like this, for many patches requires systems to be rebooted after they applied.