sure, and you could use a 10yo computer too. It sucks, but it works.
Why? Exponential gains.
how about this - a network switch is one of the most critical components in an organization
failure rates go up as they age, as well as, not being able to handle some critical traffic or some new functionality -
just about everything can be tunnelled, so why not just use old switches and virtualize everything? because that is dum.
Next - nobody was ever fired because a brand new switch failed after a month and the service people were there that afternoon because of the SLA.
(hot and cold spares might be a good idea too - depending on how critical any one piece is)
compare this to the manager who said our old switches will do, i can support them. And they STB in some cascading, meltdown fashion.
Debugging and recover efforts might take an extended amount of time. I wouldn't want to be standing there.
It isn't about the cost of the new switches, or the perceived obsolescence of the old ones.
It is the cost of keeping the business up and running, and supporting any approved edge case,
and being prepared for next year. People don't do well when the technical infrastructure is down,
even tho they get paid to think, most can't get anything done.
couple other adages:
- When making a big decision, hire a consulting company to tell you what you already know.
- Nobody was ever fired for hiring IBM.