The sorry state of the blockchain

blog.luisivan.net

8 points·luisivan··3 comments

I respectfully disagree.

I think there is a big role for open blockchains, but if 20 traders want to sit under a buttonwood tree and run their own blockchain there is no problem with that.

One of the reasons why blockchains took so long to arrive is that from a conventional "distributed systems" point of view they are a nonstarter. If you double the number of nodes, you don't increase the capacity of the system to do work. You increase the security, you avoid the "long tail" latency problem, and you can execute "smart contracts" without adding communications overhead, but the economics are not favorable to execute your "smart contract" on 10,000 nodes at the same time.

If you really want to crush permissioned blockchains you need to figure out some way to keep the security benefits of blockchains but shard them in such a way that you can add more nodes and increase the workload capacity. If you don't do that, permissioned blockchains will win for most applications./

Thanks for the comment, I understand the problem, but honestly I think you can just have a distributed database with Paxos/Raft and some code in the machines, and that's it. And it's OK! Just that it's not a blockchain, and calling it that way creates this crazy confusion we're embedded in.

You can also take the same software that you'd use in a public blockchain and put a lock on it. For instance, the Consensys folks set things like that up for banks.

Reply on news.ycombinator.com