Q: Could/Should it be adopted as an S.O.P. that machines in a server capacity (e.g. file server, web server, authentication server, Intranet server, etc.) should be configured with some level of RAID?
A: Hello, and welcome once again to Benjy's House of Incredibly Broad Generalizations!
Today's Incredibly Broad Generalization:
Question any standard intended to save you from the burden of having to think about what you're doing. It can offer you a place to start, or a default to use, but it's important for people to understand what they're doing regardless, because
1) All the moments of existence are as snowflakes, no two being alike.
I once walked around a data center and counted twelve TLD892 tape stackers which had been racked and never used. Seems someone decided that new systems should be configured with some kind of backup system. Unfortunately, nobody included a step in the process which said, "Now take a moment and consider: Are the Suspiciously-Merry Backup Fairies going to be implementing a backup schedule, tape rotation, off-site agreement, and testing procedure? OR ARE YOU GONNA HAVE TO DO IT YOURSELF, MR SMARTYPANTS ENGINEER?"
2) Shit happens.
(therefore) there will be much gnashing of teeth, and many cries of "But it worked when Joe did it!"
Obviously the SMBF weren't available for contract that week. Nobody ever actually did anything to make use of the tape changers, each of which cost something like four times the value of my beloved departed 1990 Jeep Wrangler, "Kirby". For some reason I'm reminded of the cargo cults of the Pacific islanders which built landing strips in an effort to get the airplanes to return with more whiskey and syphilis.
Having just spent six months of my life that I'll never get back on a network backup solution, I was a bit miffed over the continued purchase of unused, unnecessary, uneconomical hardware. I mean, I like green LEDs as much as the next guy, but there's cheaper ways to make your racks look festive.
I indicated my displeasure over this situation. After the wounded were airlifted and the eyewitnesses sedated, we found that backup considerations were now being evaluated with somewhat more diligence and on a case-by-case basis.
Rather than specify that RAID be used as an SOP, maybe you could instead write up an engineering process that includes a section on engineering for availability/fault-tolerance/recovery/blame-avoidance that explains what sorts of hardware are approved for purchase by the Evil Procurement Lemurs and under what conditions one would want such hardware.
Section II. Blunt instruments
Blunt instruments are indicated for use in close-quarters combat in enclosed spaces, densely populated areas, or when a swift demise is too good for the bastards.
[ ] Olympia Industrial 64-436 1/2" x 12" Gooseneck Wrecking Bar
Preferred for low-budget cosh and dislocation needs.
[ ] Olympia Industrial 14" Heavy Duty Pipe Wrench
Best for kneecaps and elbows. A traditional favorite; the Dom Perignon of tolchockers.
[ ] Death Stick EX-9 Exhumer
Really scary-looking. Also has bottle opener. Ideal for interacting with NT user base, a task best approached through the numbing haze of a few Coronas.
The danger, of course, is that if folks are given a menu of multiple-choice options, they'll engineer systems with about as much consideration as they apply to more mundane decisions, such as whether to Super-Size their fries. But sometimes this is preferable to giving them the opportunity to be creative in expensively non-functional ways.
Best of all would be a general-purpose document that outlines the issues to be explored and the decisions to be made in engineering systems, starting with "What business need exists to be fulfilled?" and ending somewhere around "Bondi Blue or Raspberry?" I'd write such a document, but the inspirational voices in my head have moved on to stain removal tips (White vinegar is your friend! Dab, don't scrub!) so I'll leave off here.