jackrabbit-oak-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thomas Mueller <muel...@adobe.com>
Subject Re: MongoMK^2 design proposal
Date Thu, 14 Feb 2013 08:52:31 GMT
Hi,

>I'm a bit skeptical if the added complexity is worth the disk space
>saving.

As for disk space: when storing data as JSON (as I guess 99% of all users
of MongoDB do), MongoDB compresses the data as well. Currently the format
seems to be BSON, and there are plans to Snappy / LZO / LZ4. See also
https://jira.mongodb.org/browse/SERVER-164

As for in-memory: it depends on what level we cache, on whether we cache
Java objects (node objects / property objects) or a lower level
representation. Caching a segment might not be efficient if only a small
subset of the segment is accessed. Caching on a finer grained level (just
nodes) might be more efficient.

So I guess disk space and memory usage is just "different", but it's not
so easy to say whether it's "better" or "worse".

Regards,
Thomas


Mime
View raw message