When our database reaches the 10-100TB range, we may need to consider non-relational databases. Relational databases like MySQL tend to heavily rely on random access, which is governed by the slow disk seek rates. On the other hand, non-relational database pioneered by Google's Map/Reduce framework operate at much faster disk transfer rates. Open source solutions incorporating these ideas exist, such as Yahoo's Hadoop and CouchDB.
Are there currently (or plans for) any mySQL engines with this type of database?
