Sirix is a temporal storage system effectively and efficiently storing snapshots of time varying (currently semi-structured) data taking full advantage of flash based drives as for instance SSDs. We not only provide several ways of navigating the tree-structure of a single revision, we also support navigation in time. Furthermore we provide a novel storage technique called sliding snapshot to circumvent intermittant full dump snapshots to fast track their in-memory reconstruction and thus we avoiding write peaks and having to read long chains of page fragments/increments/deltas. Sirix uses copy-on-write (COW) as well as an append-only storage making it an ideal candidate for flash based drives while not dropping support for erstwhile disks.
Sirix is a storage system, which brings versioning to a sub-file granular level while taking full advantage of flash based drives as for instance SSDs. As such per revision as well as per page deltas are stored. Currently, we provide several APIs which are layered as for instance a RESTful, asynchronous HTTP-API, an XQuery-API...
SirixDB (Open Source) - A Time Machine For Your DataStoring and querying of your temporal dataApr 2018