I feel like I read somewhere back about blockchain databases but I can't seem to find anything too solid on the topic. This paper called Blockchain Databases doesn't offer much in terms of description. It is just a few pages, mentioning "Hashchains" and that "Blockchain databases are structured as mutual distributed ledgers (MDLs)", which doesn't mean much.
This one, vChain: Enabling Verifiable Boolean Range Queries over Blockchain Databases, mentions "From the database perspective, a blockchain can be seen as a database storing a large collection of timestamped data records". They also mentioned FlureeDB, working on a SQL blockchain database, and BigchainDB. Both of those projects have little in terms of describing implementation detail on how exactly the "blockchain technology" interfaces with "database". Further, they mention:
The data stored in the blockchain can be modeled as a sequence of blocks of temporal objects {o1, o2, · · · , on }. Each object oi is represented by ⟨ti ,Vi ,Wi⟩, where ti is the timestamp of the object, Vi is a multi-dimensional vector that represents one or more numerical attributes, and Wi is a set-valued attribute. To enable verifiable query processing, an authenticated data structure (ADS) is constructed and embedded into each block by the miners (to be detailed in Sections 5-7). We consider two forms of Boolean range queries: (historical) time-window queries and subscription queries.
In the FlureeDB paper, it states:
FlureeDB is a protocol token whose blockchain stores a series of point-in-time facts as atomic units we call Flakes. Every Flake is — and forever will be — unique. Flakes can be thought of as events asserting (or retracting) a fact at a particular point in time. A block is a collection of Flakes as a single atomic transaction, and represents each state change of the blockchain.
...
Every database is a separate blockchain, and groups of databases are placed into shards
...
When submitting a transaction, a user will spend their own tokens unless the database has established rules that allow it to spend tokens that it holds. FlureeDB allows a rich permission model that can control exactly what pieces of data can be modified by whom and also which transactions must be paid for by the transacting user. [NOT DESCRIBED]
Still after at least browsing through these papers I don't see any descriptions of how the data storage of database records works in a blockchain / Merkle tree.
My question is specifically, how blockchain applies at the level of storage. I would like to know how the concept of a table, record, column, and value, and relations/links between data, are generally accounted for in a blockchain. That's pretty much the only thing I don't see in any of these resources, how the data model is implemented in the blockchain.
Maybe this Filecoin paper will be helpful.
