Integration of an in-memory database

[elfring]

The JkDefrag library uses a tree to process the data structures on partitions. I suggest to reuse software for in-memory databases. I imagine that the optimisation process can benefit from enhanced filtering capabilities by file system queries in SQL.

:?: How do you think about to apply components like MonetDB or FastDB for defragmentation purposes?

[jeroen]

:?: How do you think about to apply components like MonetDB or FastDB for defragmentation purposes?

Thank you for your suggestion, but no, absolutely not. The balanced tree that JkDefrag is using right now is the fastest possible way, and uses as little memory as possible. An in-memory database will make it an order of magnitude slower, and will bloat memory usage to totally unuseable numbers.

[elfring]

Would anybody be interested to see if software like the multidimensional hierarchical toolkit can enhance JkDefrag in meaningful ways for more powerful file system queries?