In his first announcement a separate NoSQL database Oracle revealed few
details, so hang one thing remained: The basis of it all is the Berkeley
DB, an established long-standing key-value store.
The impression that they had stuck to just another label that would not
leave, and Oracle is now clear that his NoSQL database offers a lot more
than the file-oriented BerkeleyDB. They bring among other replication,
dynamic partitioning of the data ("sharding"), load balancing,
monitoring, and monitoring functions as well as backups across multiple
nodes.
Also the interface is different from BerkeleyDB. Remains exactly as it
looks, though still unclear. Oracle notes, however, that the Java API
for key-value pairs CRUD functions (create, read, update, delete) and
"iteration" offer. The latter could stand for cursor support. There are
also major and minor keys, and all pairs are a common master key in the
same replication group.
In addition, the NoSQL abstracting from the underlying database
replication technology. This user would have to worry about the machines
that act as master or slave.
No comments:
Post a Comment