Why do we still care about RAID levels?

[Jeggs101]

Philosophical question I'd imagine at this point. Why do we still care about RAID levels? I mean it's a tech/ schema from single core processor days. You'd think software would just automatically take resources available and manage redundancy and performance. If I've got two drives 1/2 full, why do I choose RAID 1 or 0? Why can't the OS just balance with input of 20% free space then maintain redundancy as possible?

 

[Patriot]

Philosophical question I'd imagine at this point. Why do we still care about RAID levels? I mean it's a tech/ schema from single core processor days. You'd think software would just automatically take resources available and manage redundancy and performance. If I've got two drives 1/2 full, why do I choose RAID 1 or 0? Why can't the OS just balance with input of 20% free space then maintain redundancy as possible?

And you have just created storage spaces... or snap raid or.
Hardware raid as a whole tends to have better performance and stability than software solutions.

Scale out... and you tend to see software solutions and HBAs. (talking multiserver PB level)

 

[Jeggs101]

Even with SR or SS you are explicitly setting redundancy levels. what do I care if I have a 400 GB drive with 100GB of data on it and there are two copies kept instead of one for speed/ anti file corruption purposes? I'd imagine it makes drive space available calculations harder but those queries are not that bad from a number of times run daily planning perspective.

Even most scale out like scaleio doesn't do it this easily.