[andreskytt]

AFAIK Indian telecoms are using blockchain tech to maintain a distributed database of do-not-call numbers. Lag of up to 24 hours is ok and the relatively small number of parties is quite trustworthy. Basically they just use the bc for consensus and mirror it to a rdbms for read queries. That’s the only bc implementation I know that is both in production and is not speculation- or crime-fed.

[ChymeraXYZ]

Hey, I have a few serious questions, as I have really been trying to understand the benefits of using blockchain over "classic" stuff. They may sound like I'm questioning things, and that's because I am but with the goal to further my understanding.

> Lag of up to 24 hours is ok

So you could use any eventually consistent DB? Or a pub-sub type queue that you process at your leisure?

> and the relatively small number of parties is quite trustworthy.

I'm assuming this parties need to apply/be invited? How is that different than needing an API key?

> Basically they just use the bc for consensus

Consensus of what exactly? AFAIU the consensus is about the state of the data on the blockchain, with sanity checking against itself, not about if many parties agree if "this number should be blocked"

> and mirror it to a rdbms for read queries.

I was always under the impression that it's quite fast to read the data off the blockchain if you are a node. Am I misunderstanding something or is what they are doing just overkill?

[edgyquant]

It will never be as quick to loop through a giant blockchain to generate a dataset and then do something with it as it is to run queries in a highly optimized database that mirrors the end result of looping through that ledger. Blockchains are kind of like git and you need to move through each commit before you can get a snapshot of the current data.

[tromp]

Some blockchains are far more optimized than that. Like Mimblewimble blockchains in which all spent outputs can be permanently eliminated without hampering verifiability of the UTXO set.

[lmm]

Blockchain in the sense of proof-of-work mining? A Merkle tree might make sense, signed updates might make sense, but I don't see that mining adds value in that use case.

[tromp]

PoW, besides being the simplest way to reach consensus, also allow for a fair distribution of new coins.

[lmm]

Huh? How does distributing new coins help with maintaining the do-not-call list?