The clarity in that explanation was almost ethereal. I’ve been wanting to and I’m going to learn the inner workings of Cardano better, but I feel like this video gave me an invaluable new layer of knowledge to what I already knew. Keep up the good work man.
I've looked into ouroboros Leios since the research paper came out but it looks like there's still a lot more research needed for how it will be implemented and there are still many unanswered questions around it so I think Leios is still multiple years away. I might make a video when it's more concrete. From my understanding, essentially validators will run a VRF for producing input blocks which occur much more often and are blocks that contain transactions. They will also run a VRF to produce reference blocks which are blocks that reference the hash of many input blocks. They also run another VRF to vote on if a reference block and all the input blocks it references are available to be downloaded and there is no information missing. After this, it's less clear on how it will work. It seems like they will keep the current ouroboros but instead of block winners including transactions in their blocks, they include references to one or many reference blocks that have passed the threshold of data being available. I'm not sure if the current consensus is fast enough for this as a reorg would mean potentially re-running a bunch of blocks so nodes would need to have extra unused capacity for these situations which is against what Leios is aiming to do in utilising more of the nodes capacity. There's still a lot of uncertainty on how it will actually be implemented.
This is insanely explanatory. Love your videos. Can you do a breakdown on Everscale (the tech is based on Telegram’s whitepaper)? I think it is one of the best techs in crypto and has some smart implementations, similar to Cardano’s short and long forks handling
doesnt the plenitude rule just delay the long range attack? if the dishonest can catch up with the honest chain, wouldnt it also sooner or later become more dense due to the main chain not producing blocks as you explain due to validators going offline or network delays?
The plenitude rule only looks at the density immediately after the fork so even if the attackers chain becomes more dense later on, it doesn't matter. The plenitude rule isn't a complete defence against long range attacks as an attacker could try buying old staking keys which may be cheap and may be able to get enough stake to create a very dense chain immediately after the fork but that would need like 90+% of the stake so would be very difficult. I believe Cardano may use Key Evolving Signatures as well where staking keys are deleted after they are used. There's no way to enforce nodes to delete their keys but in the default software they do so it's likely impossible to currently perform a long range attack on Cardano as old staking keys have likely been deleted
@@altexplainer Yea, but why cant the dishonest chain be denser? Without the plenitude rule, the dishonest chain will broadcast the chain after it becomes the longest, but with the plenitude rule, they just have to broadcast it when it become the longest and more dense. no?
The plenitude rule only looks at the density of blocks immediately after the fork. If an attacker has the staking keys for 60% of the stake from a year ago and starts creating their alternate fork. At the initial point where the chains fork the attackers chain will only be 60% as dense as the honest chain (assuming 100% of stake was creating blocks for the honest chain at the time). Over time the attackers chain will become more dense but the plenitude rule is only looking at the initial point where the chain forks and at that point the attackers chain would only be 60% as dense
Thank you, this was very helpful. How can SPOs decide which fork(s) they operate on? Isn't it that basically just cardano-node is running and doing its thing? I assume blocks in non-winning chains do not get rewards, right? I just started running an block producer on testnet, and the recent issues around version 1.35.x of cardano-node got me motivated to understand by heart what is happening behind the scenes.
The node software will handle choosing forks for you and will choose the longest fork. (Or if they are the same length, the side of the fork where the block has a lower VRF number) If you want to choose the fork yourself, you will have to modify the node software which very few people will do Yes that's correct, only blocks in the winning chain get rewards Nice, running a node and trying to understand what the people on discord are saying when there's a problem is a good way to learn
NiPoPow looks interesting. I'm planning to make some video's on zero knowledge proofs so maybe I'll do a video on Ergo after them as they look like they use similar concepts
Great explanation. The thing I don't understand is where do block rewards come from after the inflation rate reaches zero. We're down to 4% already. Shouldn't fees contribute to staking rewards for delgators?
Some of the fees go to the treasury and my understanding is the rest goes back into validator rewards. In the future if cardano is successful it is reasonable to believe that a much higher number of transactions will take place hence a larger base for validator rewards. Side chains will also provide a source of valudator rewards that will encourage ADA holders to stake.
The clarity in that explanation was almost ethereal. I’ve been wanting to and I’m going to learn the inner workings of Cardano better, but I feel like this video gave me an invaluable new layer of knowledge to what I already knew. Keep up the good work man.
You’re the GOAT at explaining crypto concepts
Brilliant explanation, particularly liked the detail on attacks and mitigations. Nice work!
You deserve 1 million subscribers
The amoung of knowledge and information is insane. Thankyou for your amazing content for a teen to learn about how cardano actually works!
More Cardano content please 🙏
Great explanation
awesome explanation. kudos
6:28 - time (chronos)
9:04 - long range attacks
10:28 - plenitude rule
11:14 - unsolved: appending to both sides of a fork
Super content. Thank you very much
would be nice a video about ouraborus leios (imput endorsers) and it's potential
I've looked into ouroboros Leios since the research paper came out but it looks like there's still a lot more research needed for how it will be implemented and there are still many unanswered questions around it so I think Leios is still multiple years away. I might make a video when it's more concrete.
From my understanding, essentially validators will run a VRF for producing input blocks which occur much more often and are blocks that contain transactions. They will also run a VRF to produce reference blocks which are blocks that reference the hash of many input blocks. They also run another VRF to vote on if a reference block and all the input blocks it references are available to be downloaded and there is no information missing.
After this, it's less clear on how it will work. It seems like they will keep the current ouroboros but instead of block winners including transactions in their blocks, they include references to one or many reference blocks that have passed the threshold of data being available. I'm not sure if the current consensus is fast enough for this as a reorg would mean potentially re-running a bunch of blocks so nodes would need to have extra unused capacity for these situations which is against what Leios is aiming to do in utilising more of the nodes capacity. There's still a lot of uncertainty on how it will actually be implemented.
thank you, very helpful video! :D
Great video!
Glad you enjoyed it!
Hello, may I know which version of Ouroboros have you explained in this video?
Hey, it's on Ouroboro Praos which I believe is the current version
This is insanely explanatory. Love your videos. Can you do a breakdown on Everscale (the tech is based on Telegram’s whitepaper)? I think it is one of the best techs in crypto and has some smart implementations, similar to Cardano’s short and long forks handling
Thanks, I'll add Everscale to my list of projects to look at. It's probably going to be a while though
doesnt the plenitude rule just delay the long range attack? if the dishonest can catch up with the honest chain, wouldnt it also sooner or later become more dense due to the main chain not producing blocks as you explain due to validators going offline or network delays?
The plenitude rule only looks at the density immediately after the fork so even if the attackers chain becomes more dense later on, it doesn't matter.
The plenitude rule isn't a complete defence against long range attacks as an attacker could try buying old staking keys which may be cheap and may be able to get enough stake to create a very dense chain immediately after the fork but that would need like 90+% of the stake so would be very difficult.
I believe Cardano may use Key Evolving Signatures as well where staking keys are deleted after they are used. There's no way to enforce nodes to delete their keys but in the default software they do so it's likely impossible to currently perform a long range attack on Cardano as old staking keys have likely been deleted
@@altexplainer Yea, but why cant the dishonest chain be denser? Without the plenitude rule, the dishonest chain will broadcast the chain after it becomes the longest, but with the plenitude rule, they just have to broadcast it when it become the longest and more dense. no?
The plenitude rule only looks at the density of blocks immediately after the fork. If an attacker has the staking keys for 60% of the stake from a year ago and starts creating their alternate fork. At the initial point where the chains fork the attackers chain will only be 60% as dense as the honest chain (assuming 100% of stake was creating blocks for the honest chain at the time). Over time the attackers chain will become more dense but the plenitude rule is only looking at the initial point where the chain forks and at that point the attackers chain would only be 60% as dense
@@altexplainer I see, tyvm. This make sense.
Thank you, this was very helpful. How can SPOs decide which fork(s) they operate on? Isn't it that basically just cardano-node is running and doing its thing? I assume blocks in non-winning chains do not get rewards, right?
I just started running an block producer on testnet, and the recent issues around version 1.35.x of cardano-node got me motivated to understand by heart what is happening behind the scenes.
The node software will handle choosing forks for you and will choose the longest fork. (Or if they are the same length, the side of the fork where the block has a lower VRF number)
If you want to choose the fork yourself, you will have to modify the node software which very few people will do
Yes that's correct, only blocks in the winning chain get rewards
Nice, running a node and trying to understand what the people on discord are saying when there's a problem is a good way to learn
Try making a video on Ergo, which is a new model on proof-of-work blockchain
NiPoPow looks interesting. I'm planning to make some video's on zero knowledge proofs so maybe I'll do a video on Ergo after them as they look like they use similar concepts
Great explanation. The thing I don't understand is where do block rewards come from after the inflation rate reaches zero. We're down to 4% already. Shouldn't fees contribute to staking rewards for delgators?
Yes all the fees from an epoch are added together and distributed among the delegator nodes as a staking reward
Some of the fees go to the treasury and my understanding is the rest goes back into validator rewards. In the future if cardano is successful it is reasonable to believe that a much higher number of transactions will take place hence a larger base for validator rewards. Side chains will also provide a source of valudator rewards that will encourage ADA holders to stake.
Wow
8:40