mleku on Nostr: so, today i learned about #nostrocket and i already knew about #negentropy and i'm ...
so, today i learned about #nostrocket
and i already knew about #negentropy
and i'm like, ok, so people kinda get that some kind of loose distributed consensus is a good idea for nostr
i'm currently building a relay with an interface to a "sMaRt CoNtRaCt" distributed database engine
i think there is a clear case for some types of events to be replicated fast over blockchain style byzantine fault tolerant database clusters, many, the more the merrier
i think that people in #nostr with their #bitcoin only leanings are pushing too hard to exclude shitcoiners from nostr when the single thing that their silly systems do that is actually relevant to us, is something bitcoin can't do
and we don't want it to:
storing bulk data
people are a bit confused about this
Practical Byzantine Fault Tolerant (pBFT) database replication protocol was invented in 1999
it enables a cluster of replicas of a database to stay in sync to a pretty tight time window, according to some simple limitations:
- each extra node in the network needs an extra message at each step in the protocol for agreeing on the state change
- so long as 2/3 of the nodes agree it doesn't matter whether the other 1/3 are dead or malicious
practical implementations as tested by the Cosmos project show that about a maximum of 99 replicas can be added to a cluster before it starts to get bogged down by waiting for messages
in contrast, the bitcoin ledger can have literally millions of replicas because the difficulty adjustment of the nakamoto consensus doesn't need to account for the voting process of pBFT protocol messages
the block hash itself is the permission to publish a new block
it solves the problem of latency but the trade-off is that the more mining resources there is available in the world, that aren't mining on a given database, the less the security of the protocol
that is, it's very difficult to make a proof computation that is specialised enough that nobody with the capacity to do it is doing anything else except mining on that database
this is why the SHA256 hash is actually perfect for the job, and why there can only be one monetary database for our planet, and probably at least solar system
i'm not sure if y'all grasp that i'm "working with the enemy" on my relay project, but i'm not considering them to be enemies except as far as they perpetuate the meme that staking has anything to do with security and is anything more than a mechanism to contain the number of pBFT replicas involved in the consensus
y'all need to get used to the idea that shitcoiners are coming to #nostr
they are coming whether you like it or not
but i'm just here to tell you that the pBFT protocol is a good idea for scaling nostr, and it doesn't need to be centralised, i can imagine hundreds of pBFT clusters each with different implementations each replicating a relay for the network as a whole
even #nostrocket uses a staking scheme (peg based, like liquid)
:facepalm:
when i have the resources, i'm going to build a clean, pure human voting consensus based pBFT replication based relay for nostr databases
i think that social acceptance is a more powerful regulator than fucking stake, basically. you can make a replica of our relay cluster, because you seem decent, and we keep you in the group because you stay decent
#devdiary
and i already knew about #negentropy
and i'm like, ok, so people kinda get that some kind of loose distributed consensus is a good idea for nostr
i'm currently building a relay with an interface to a "sMaRt CoNtRaCt" distributed database engine
i think there is a clear case for some types of events to be replicated fast over blockchain style byzantine fault tolerant database clusters, many, the more the merrier
i think that people in #nostr with their #bitcoin only leanings are pushing too hard to exclude shitcoiners from nostr when the single thing that their silly systems do that is actually relevant to us, is something bitcoin can't do
and we don't want it to:
storing bulk data
people are a bit confused about this
Practical Byzantine Fault Tolerant (pBFT) database replication protocol was invented in 1999
it enables a cluster of replicas of a database to stay in sync to a pretty tight time window, according to some simple limitations:
- each extra node in the network needs an extra message at each step in the protocol for agreeing on the state change
- so long as 2/3 of the nodes agree it doesn't matter whether the other 1/3 are dead or malicious
practical implementations as tested by the Cosmos project show that about a maximum of 99 replicas can be added to a cluster before it starts to get bogged down by waiting for messages
in contrast, the bitcoin ledger can have literally millions of replicas because the difficulty adjustment of the nakamoto consensus doesn't need to account for the voting process of pBFT protocol messages
the block hash itself is the permission to publish a new block
it solves the problem of latency but the trade-off is that the more mining resources there is available in the world, that aren't mining on a given database, the less the security of the protocol
that is, it's very difficult to make a proof computation that is specialised enough that nobody with the capacity to do it is doing anything else except mining on that database
this is why the SHA256 hash is actually perfect for the job, and why there can only be one monetary database for our planet, and probably at least solar system
i'm not sure if y'all grasp that i'm "working with the enemy" on my relay project, but i'm not considering them to be enemies except as far as they perpetuate the meme that staking has anything to do with security and is anything more than a mechanism to contain the number of pBFT replicas involved in the consensus
y'all need to get used to the idea that shitcoiners are coming to #nostr
they are coming whether you like it or not
but i'm just here to tell you that the pBFT protocol is a good idea for scaling nostr, and it doesn't need to be centralised, i can imagine hundreds of pBFT clusters each with different implementations each replicating a relay for the network as a whole
even #nostrocket uses a staking scheme (peg based, like liquid)
:facepalm:
when i have the resources, i'm going to build a clean, pure human voting consensus based pBFT replication based relay for nostr databases
i think that social acceptance is a more powerful regulator than fucking stake, basically. you can make a replica of our relay cluster, because you seem decent, and we keep you in the group because you stay decent
#devdiary