Hacker News new | past | comments | ask | show | jobs | submit login

It's not quite a chicken and egg problem. Features like ZNS come into existence in the first place because they are desired by the hyperscale cloud providers who control their entire stack and are willing to sacrifice compatibility for small efficiency gains that matter at large scale.

The problem for the rest of the market is that the man-hours to rewrite your software stack to work with a different storage interface that allows eg. a 2% reduction in write amplification isn't worthwhile if you only have a fleet of a few thousand drives to worry about. There's minimal trickling down because the benefits are small and the compatibility costs are non-zero.

Even simple stuff like switching to shipping drives with a 4kB LBA size by default has very little performance impact (since drives are tracking things with 4kB granularity either way) and would be problematic for customers that want to apply a 512B disk image. The downsides of switching are small enough that they could easily be tolerated for the sake of a significant improvement, but for most of the market the potential improvement is truly insignificant. (And of course, fancy features are a market segmentation tool for drive vendors.)




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: