Hi,
I would like to get your opinion on providing Hibernate-backed JPA
implementation for non-relational storage.
Given rising popularity and variety of key-value stores and the fact
that often they provide subset of functionality found in relational
databases, it might make sense (especially for web-scale companies that
most likely will be using number of storage technologies) to leverage
Hibernate model for key-value stores. Perhaps only a subset of features
and subset of the querying capabilities.
As relations between entities in key-value stores are done on client
side, the possible implementation could include some form of relational
engine (similar to what Carbonado[1] and Oracle's DPL[2] do for
BerkeleyDB) to provide (partial) support for query language. Providing
caching, entity lifecycle management, validation and even transactions
should be rather easy as those are orthogonal concerns.
WDYT? Do you see Hibernate/JPA catering for scenarios that leverage
key-value stores? Do you see any major impediments either in Hibernate
implementation or JPA philosophy that would make such extension not
worth the effort?
Cheers,
Tomasz
ps: sample project along similar lines of thinking:
http://code.google.com/p/simplejpa/
[1] http://carbonado.sourceforge.net/
[2] http://www.oracle.com/database/berkeley-db/je/index.html
_______________________________________________
hibernate-dev mailing list
hibernate-dev@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hibernate-dev
