Western Digital announces 26TB hard drives and 15TB server SSDs

arstechnica.com

9 points·lisper··4 comments

I can still remember my first day at Google in spring of 2000. A small crowd was gathered around a newly built server which had been stuffed to the gills with 25 brand new 40GB hard drives, to which Google had just acquired an exclusive deal. I remember thinking, "Wow, that's a terabyte", an inconceivably large number back then. What would you ever do with that much storage?

Today a terabyte costs $99 and fits in my pocket.

Unfortunately storage cost reduction has slowed / halted and we have no concrete information to suggest things will drastically improve. The new 22TB CMR HDD uses 10 Platter, i.e there is no increase in storage density per platter. NAND is similar as well with Multiple String stacking.

I'm trying to build a low cost NAS for my paltry needs, but the price per _drive_ doesn't seem to be moving. Every time a larger size is introduced, I scoff at the price for something I don't need, but the price of a 4TB doesn't go anywhere. I bought four drives on eBay, they were 80% of retail price, and all four were dead. It's pretty frustrating but I don't want to spend so much money that I don't have.

In combination with the pricing... I feel like mechanical drives in particular are getting ridiculously large, too.

I have a large collection of 5TB drives, and it's honestly already at the point where instead of going larger, I'd simply get more drives.

Peaking at ~200MB/s writes means that making practical use of this much data is getting more and more specific (eg: archival). It can make more sense to build an array of smaller drives to get the same capacity but actually useful performance.

Reply on news.ycombinator.com