In IT more hardware is redundancy, but in aviation a liability
In IT, having redundancy and diversity is low cost and no liability, but high gains in service reliability.
In aviation, redundancy and diversity is high cost and a liability, for no increase in reliability of the service.
The long answer
The big difference is that in case of a hardware failure in IT, you just lose a bit of capacity. The service stays up, the customers do not even notice, you revenue stays intact. In IT, a failure is all right, and is allowed to happen.
In aviation, failure means you lose the entire service. All operations are centered on the philosophy that failure is not an option, in the sense: we can deal with it... but it costs a lot and we really do not want to do it.
Let us go through all the consequences of a failure...
Injury, loss of life
- IT: a server blade goes down. So what? No-one is injured, no-one dies.
- Aviation: An engine goes to pieces. That. Is. Not. Good. Losing/injuring colleagues and/or clients on the job must not happen.
Disruption to service, loss of revenue
IT: a server blade goes down. So what? The performance of the service is slightly degraded. But customers are not turned away. You do not lose revenue over it.
Aviation: an engine comes apart. You are now rapidly losing money, because as of that moment that aircraft is a cost, not a source of revenue.
The key difference here is that in IT you are only losing a percentage of the service. In aviation, as soon as an engine failure happens, you lose the entire service. This means more engines are a liability, because they mean more possible points of failure.
First the flight in question is instantly cancelled/interrupted, because you do not keep flying as if nothing happened. You are setting the aircraft down on the ground and you are doing it now. This means accommodating all the passengers, either hosting them in a hotel or rescheduling them on other flights... fares you have to pick up the tab for.
Second: the aircraft is now out of commission for unscheduled repairs. This is a big deal because airliners earn all their revenue while flying; they earn nothing while standing around in a repair-shop. An airliner spends about 2/3 of its service life in the air. That is to say: 16 out of 24 hours every day, for 20-30 years, an airliner is supposed to be in the air making money.
Third, you can bet your bottoms that someone had their phone up — even though you told them to turn it off — and filmed the whole thing happening. Then they call their tabloid and email over the link to their YouTube clip of it happening. Your airline's logo is now all over the evening news, your failure published nation- and world-wide. This means loss of customer confidence, which means more loss of revenue.
Replacing the busted hardware
- IT: the cost of the broken hardware and the manhours needed to replace them are rarely counted in excess of 10 000 USD. Now granted this happens a whole lot more often than an aircraft engine failue, but in general, this is a very small cost.
- Aviation: The cost of an engine is in the range of 5 000 000 USD and up. A new Rolls Royce Trent 1000 is 15 000 000 USD. Add to that hundreds of manhours to get the replacement engine out of storage, ship it to the location of the wounded bird, remove the busted engine, inspect the aircraft for damage and certifiy it fit enough to fly, put in the new engine, get the aircraft back in service.
Others have already gone through this so I will just mention it briefly: maintenace costs is counted per unit in aviation. Twice as many engines = twice as much maintenace cost. Such is not the case in IT.
Also — and this is the crux of your question — you ask why in aviation one goes for commonality instead of diversity. This is because in aviation knowledge and experience of a system are commodities. Double the number of types of systems (such as different engines), and you double the amount of people you need to hire, along with double the amount of experience you need to accrue.
Also the support infrastructure needed to deal with one type of system is different from the next. Again: you multiply the maintenance cost for every type of system you add to your organisation.
The only reason airliners have not gone down to using only one engine per aircraft is for safety reasons... since the only thing more unacceptable than having one engine quit in mid flight is to have all engines quit.
In IT, having redundancy is low cost and no liability, but offers high gains in service reliability. In aviation, redundancy is high cost and a liability, for no increase in availability of the service. The same goes for diversity: there are no gains to be had from it.
So in conclusion: failure is not an option. We can deal with it but — unlike in IT — any such failure is an expensive and very disruptive event. We just do not want to have to deal with it. Since more hardware, and diversity of hardware, increases both the risk of failure, and the maintenance and support costs for it, there is nothing to be gained from diversity and redundant redundancies.