Hi,
Below are some notes of recent development in Oak. Please feel free to
comment, correct and complete as necessary.
* The segment based MicroKernel prototyping effort is coming along nicely:
° Two backends, memory and MongoDB
° Mostly functionally equivalent with other MKs, some isolated errors
still to resolve Performance (with local MongoDB backend) roughly on par
with the H2 MK, with plenty of optimizations yet to be implemented
° Can already be used in a clustered scenario, the main current
concurrency limitation is the optimistic lock on the root journal
° See also https://issues.apache.org/jira/browse/OAK-593
* Progress of the new MongoMK prototype:
° Create and modify features so far
° Read (absolutely no caching so far)
° Performance roughly on par with the H2 MK
° Runs some of the oak-jcr tests
° See also https://issues.apache.org/jira/browse/OAK-619
* Some improvements to the read cache in oak-core significantly reduce
the number of round trips to the Microkernel after write operations.
Further improvements on the horizon. See
https://issues.apache.org/jira/browse/OAK-591
* Solr indexer:
° Working IndexHook and QueryIndex implementations
° Pluggable Solr providers and indexing configurations to allow custom
installations
° See also https://issues.apache.org/jira/browse/OAK-307
* Improved search API:
° Improve / differentiate the concept of query cost
° Abstract the concept of execution plan
° Take into account sorting of results in execution plans / cost
calculations
Michael
