if anyone is interested, the implementation is available as Open Source software: https://github.com/sirixdb/sirix
I recently wrote quiet some documentation at https://sirix.io regarding use cases, the APIs, put up a community forum and wrote a complete native JSON storage (besides XML) :-)
The system greatly advanced since Marc's Ph.D. thesis (now, as mentioned above it stores also JSON besides XML in a tailored binary format, I've added sophisticated query capabilities through Brackit (https://brackit.org) and temporal enhancements in Brackit (https://github.com/sirixdb/brackit), added versioned index-structures, diffing capabilities to import differences between XML-documents or to diff revisions once stored in Sirix (and added a currently outdated GUI with novel interactive visualizations of the encountered differences found between any two revisions).
Next steps will be adding rewriting rules for index accesses in the query compiler, and with the help of you, the Open Source community I want to publish version 1.0.0 soon. Afterwards, I'd love to look into partitioning/replication, but it's also up to you :-)
Any kind of comment, help, questions regarding the APIs or anything related are more than welcome :-) I'll contact Marc, soon that I'm still super eager to put forth his ideas (and some of mine and of the other students involved, notably Sebastian Graf, who also rewrote a lot of stuff for his Ph.D.)
More than happy to answer any questions :-)