Definitely worth standardising on equipment, I'd recommend it even for home or small business use, especially when you're just starting out. Murphy's law says if something goes down it will be important and at a critical time. By duplicating the hardware used on your production servers across test/development servers, or less critical production servers, you ensure that you have an alternative platform to move critical apps onto in the event of an outage.
However, one question I would ask though is how best to upgrade? Do you keep old servers and phase them out gradually as they die / get unsupportable / too heavily loaded? Whilst this saves money and spreads out expenditure on new servers, it also means that the new kit is almost certainly incompatible with the old.
So is the alternative that you replace everything at once, rendering all of your old servers obsolete overnight, but 100% compatible? Well, that's way more attractive from an idealistic IT support perspective, but from a business/financing point of view, it's a lot of cash to outlay at once, especially if you're a big business with a server farm.
Guess this angle makes the option of hiring hardware somewhat more attractive if more expensive in the long run, but what if you buy all of your company's servers? I think it would be difficult to present a good business case for upgrading them all at once in order to standardise, as most accountants will look at the hard, tangible asset costs rather than the extra cost in man-hours supporting high-maintenance hardware.
Food for thought...
Edited by - jonreade on 01/30/2003 08:18:27 AM