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

That doesn't seem completely fair. Much like blaming google for using normal commodity servers compared to Altavista using high end enterprise hardware. What matters is the reliability of the system, not some random part.

Backblaze is quite transparent about how they do things. They publish their drive reliability numbers (including brand/model numbers), storage pod design, and how their sharding/redundancy works.

Seems like most cloud storage vendors just say "We do object storage right handwave and we have lots of 9s". Backblaze says they shard your data into 20 pieces onto 20 servers and can recover with any 17 of those pieces. More details at https://www.backblaze.com/blog/reed-solomon/

Sure that's not enough redundancy for some, but at least you know what to expect and can plan accordingly. I've not see any other cloud vendor do that. Please post URLs for similar info from other companies.




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

Search: