Sure, I’ve run the numbers on that. Erasure coding slices data into n blocks, adds m parity blocks, and you only need any n of the n+m to reconstruct the whole set. That means you can cut raw storage by roughly the parity ratio while still keeping the data safe, because you’re not just mirroring, you’re distributing redundancy across many disks. It’s efficient, scalable, and gives you a precise hit on cost versus durability.