I propose we construct a kernel that will be acessed by
any other function of the recache.
Main functions:
* Pool of connections.
* Generic pool to intensive used objects (like Tuple,
QueryImpl).
* Store the mappings (DmlStructure).
* Administration of locks (in futures releases).
* JMX administration of pools, locks and distribution.
Logged In: YES
user_id=1277429
Interessante. Ainda estou analisando como se daria isso e
como separar o que realmente infraestrutura do recache e o
que infraestrutura de servidor de aplicao