@Ericloewe is right about things not always being set and forget. I was talking to a colleague who’s a R.F. engineer recently about spares and warranties and all that jazz. The same kinda thing applies here and with my IT experience.
There are folks who have bagillions of dollars in real operating budget, some others at least get big dumps of grant money even tho they have no operating budget. Then there’s the SMB/home market people who pick up the scraps. Kinda like gaming market benefits from engineering on ai or compute focused enterprise cards.
All (most?) of these big tech companies engineer high margin enterprise components. This is a primary objective. Think a 20,000 dollar server or SAN or Switch. The BOM of these parts are usually disproportionally lower than their retail value when compared to other market segments of similar technology.
In fairness, companies also design for low power devices like phones, even though this is a lower msrgin it’s a higher volume market. Think a tv streaming set top box or a cell phone.
Tape drives are in the former category not the later.
As mentioned above, Zip disks or usb hard drives all that jazz are in the later not the former.
Because tape libraries were designed as modular devices the idea is you keep spare parts. In enterprise land, you typically have multiple fully assembled devices with service contracts. This is expensive but so are on prem SANs. The margin on these kind of devices from both the VARs and the manufacturers are rather high, and even the suppliers can be costly because these are all custom designs.
The scrappy way to do is to keep a primary production one on contract and have a generation (or two, see
@gdreade ) older one. This way you can at least always read your old backups and ideally even write to them
In homelab or even scrappier land, you pick up some spare drives for your librery(s) for what shit goes sideways.
The cloud is there to make all this shit easier so you don’t think. But you pay for that privilege and they can hold your data hostage in a whole bunch of really crummy evil ways.
I was previously an infrastructure manager for a multimillion dollar public organization, but relied exclusively on grant buckets of money with strings attached. My operating budget and my staffing levels were garbage. I literally couldn’t afford to buy patch cables but I had millions of dollars in Cisco switches and WiFi APs.
Tape drives made more economical sense for me, because I could buy the spares I needed to survive. If I couldn’t pay my iron mountain bill I could recall my tapes. On the flip side, if I couldn’t pay my cloud hosting fee I was screwed because I couldn’t afford the egress fee without fighting with the CFO.
this same business model exists in other market segments for different reasons. In my homelab life, the cost of backing up my really important home videos and pictures and other stuff matters to me. The quality of those backups also matters to me. I have it on OneDrive, my TrueNAS(s), and I have a little server at a CoLo pretty far from me geographically.
But I wanted an archival copy. One that is easily accessible and potentially faster to restore than other backups for the most critical or urgent needs. No different than when I worked, we had different sets with different retentions and all that jazz,