Okay, so I was running some cost projections for our new app's backup strategy (think cold user-generated content, accessed maybe once a year) and my usual spreadsheet flagged a change. Google Cloud Storage Nearline's storage price just dropped from $0.010/GB to $0.005/GB in the Iowa region. That's... half. It happened like, yesterday?
This puts Nearline at half a cent per GB. AWS S3 Glacier Deep Archive is $0.00099/GB, so still cheaper on pure storage. But the minute you need to *retrieve* something, the math gets messy fast.
My immediate question: for a true cold-but-might-need-it scenario, does this price shift the calculus? Glacier's retrieval fees and speeds are a whole maze. Nearline now has a 1-3 sec typical read time and the new $0.005/GB storage cost. The retrieval cost is $0.01/GB for Nearline vs. Glacier Deep Archive's $0.025/GB (Standard retrieval tier). So you'd pay more to get your data out of Deep Archive, and wait hours.
Is the new play to use Nearline for "cold-but-not-frozen" data where the faster access is worth the slight storage premium over Glacier? Or does the Glacier Instant Retrieval tier ($0.0043/GB storage) still dominate that niche? I'm trying to map this to real user behavior—if a user requests an old file, they'll bounce if it takes hours, but seconds is fine. The cost per access event feels like the new variable to optimize.
Has anyone run the updated numbers yet, including network egress and operation costs? I feel like the breakeven point on number of retrievals just moved significantly.