Ask HN: Cheapest way to store 6TB online?
https://news.ycombinator.com/item?id=4605014What is the cheapest way to store 6TB of key-value data online?
The data is accessed via random lookups and needs to maintain <0.5s service times per lookup and can expect around 200 concurrent accesses at any time.
It's currently hosted in a small Cassandra cluster and that gives lookup times of 15ms-20ms, well within acceptable limits. However there's no reason it couldn't be moved to another indexing system (it could even be moved to a suitably tuned filesystem with little effort).