My apologies in advance if there is a more appropriate category for this discussion.
Please have a look at this list of data sources. It is a list of databases implemented in a standard called Resource Description Framework. RDF allows the world to share a web-scale database schema where shared, reused URIs replace both primary keys and table/column names. These open schema are called ontologies and they are maintained by communities of enthusiast of the field being modeled (for more info on the semantic web vision, this is a good place to start).
The problem is that our RDF databases currently live in silos, meaning they (like most applications) run in isolated processes, which means isolated state, which makes confederated query results a difficult problem to solve (no viable solutions are known AFAIK). When I started studying the EOS vision, I began to wonder if the EOS blockchain could allow you to take the persistence object of one of these databases and swap it for an EOS smart contract running on the blockchain. Suppose the database uses an rdf.db file it writes and reads the db state to and from, and uses db_file_source to manage file i/o. If I traded db_file_source with eso_source, where eos_source implemented a file interface but was reading/writing to an EOS blockchain, would that be feasible from a performance standpoint, since the focus here is on the transaction relay speed (which I haven’t been hearing much about) and not so much volume of transactions in a given time.
What I’m proposing is that all these RDF databases write their db state (perform their CRUD operations) against one common persistence object, EOS. That way, when I query one of them, that instance of the database can both lookup and reason against the “global persistence object” to derive a much richer and more disparate set of results. Only the low-level persistence object needs to run on the block chain, which would minimize the amount of effort required to port the databases over to EOS.
My question is, can EOS meet or exceed the performance of a file-backed persistence. If so, my next question is are there any here who would be willing to contribute to an open source effort to make a EOS-backed persistence for RDF databases? Or could you point me to any places I could find such persons? I suspect if EOS can handle this requirement that there are a lot of enthusiastic folks in the semantic web community who would be interested in a viable solution for confederated queries.