Re:The Future Database
by throwusawayus on Monday May 23, 2022 @07:29PM
(#31486136)
> I suspect the intention is something more like IPFS (https://ipfs.io), built on some distributed data structure like a DHT.
the company's products are all built on top of sharded mysql .. unless they plan to throw away everything they have done so far, i do not think these assumptions are correct!
- "Infinitely scalable" in the sense of the internet, I suppose? If we had a way of paying for storage and indexing in a decentralized way, then you could switch from one 'provider' to another, or host your own data. Data would not be siloed in the way it is by AWS.
- I assume "serverless" means you don't have some specific upstream server you must hit with requests. You could submit queries to a local job, which could pass them to any node in the network.
- "Fundamentally reliable" because data is replicated, and jobs can be executed by any member of the network.
- "Distributed globally" because you could store data from any internet-connected device, "available locally" because, again, you don't have some single-point-of-failure server you need to connect to, or VPN you need to join, or whatever: your computer would be a node on the network, as capable of accessing data and running jobs as any other node.
And "intelligence" and "no performance cost" are just aspirations.