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

As a differing example of scaling - Facebook works very well using MySQL at a very large scale; implying that you don't need to move to key/value stores to scale.

Both technologies have their place and reasons to exist, but it's not solely for the ability to scale.




I don't necessarily disagree with your point, but that is a poor example.

Facebook uses MySQL, largely, as a key-value store.


Do you have a citation for this? The closest I could find is a mention in a gigaom article[1] that mentions that they have some data better suited for a document store tool, but saying "there likely are unstructured or semistructured data currently in MySQL that are better suited for HBase" doesn't imply that the majority of their data is key/value based.

If you'd like another example at slightly less than Facebook size - RightNow (recently acquired by Oracle). They manage customer service for many (1,000+) different clients at huge scale: more than 300 MySQL databases spread throughout the world. If a website has a knowledge base, there's a good chance it's being managed on the backend by RightNow.

[1] http://gigaom.com/2011/12/06/facebook-shares-some-secrets-on...




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

Search: