Redundancy – The Microsoft Azure platform has an advantage in almost every scenario regarding the redundancy of its infrastructure. This would be impossible to replicate in an on-premise server environment without spending millions of dollars.
In the event of a server failure, the next server in the rack can pick up the workload without skipping a beat.
If the entire server rack fails, the workload can be sent to the next rack.
If the entire data center fails, then the workload can be shifted to the following data center, if a whole region of data centers fails, then it can be moved over to an alternate region, and finally, if, for some reason, the entire United States Azure infrastructure went offline, then your workloads can still operate in other globally available datacenters.
On-premise power outages – If a server is on-premise, you are beholden to the power it provides. All of your work stops if power is dropped (without a generator onsite).
In Azure, this is a non-issue. If the power at your facility drops, you can always do your work from home, Starbucks, the library, or anywhere that can provide an internet connection. Work doesn’t have to stop just because the power has.
On-premise Internet Outage – If the power is on but the internet is not, your remote workers cannot do their jobs. Azure solves this by, again, having more redundant internet connections that can rival most ISPs.