You remember that time that you wrote a comment and it turned into a damn essay? This is that comment.
Never done this silly "open letter" thing before.
Let's see if I get the gist of it.
https://peakd.com/hive-167922/@taskmaster4450/hive-a-resource-credit-crunch-in-2021-2022
With all due respect, I feel like the way resource credits and Hive bottlenecks are described here is somewhat misleading. Hive coins and Hive Power have no affect on our throughput. Resource Credits are a completely separate resource from everything, and even they do not determine the throughput, volume, or limitations of the network. RCs simply exist as a way to limit accounts from filling up the blocks.
At the core of this issue, the only thing that matters is how much information we can store on a single block (the blocksize limit, if you will). It is the same limitation that both Bitcoin and Ethereum have. Hive is no different, except that we pay 20 nodes a lot of money instead of spreading that out to a ton of miners. We centralized our chain so that we could ensure high throughput. This is the classic scaling issue. Do you want more decentralization or more scaling? Sacrifice one for the other.
Hive blocks are currently 65KB MAX for the entire network.
You'll recall that I filled an entire block myself just to test it.
https://peakd.com/steemdev/@edicted/65-kb-max-length-post-keywords-javascript-steem-api
In this context, it's much easier to see how we could run out of room and fail to scale at any moment. We can only save a MAXIMUM of 22KB per second (1.8GB/day) to the database. This leads to the exact issue that caused Bitcoin and Bitcoin Cash to fork:
Why don't we just raise the blocksize?
Easy.
Well, it's not that easy, because Hive node operators are already complaining about how expensive everything is, even though we are constantly developing scaling solutions like MIRA and the like. Imagine what it would be like if our blocks were actually being filled to the brim and 1.8GB of raw data was being added to the Hive blockchain every day.
For each bit of raw block data added to the chain, much more indexes and fast-access data and tables are stored in expensive ram on the nodes. Even if we move that data to SSD storage, it becomes much slower and there is still a cost to keep it there.
Which is why we have resource credits: to stop the blocks from filling up. However, if the blocks fill up anyway we have to raise RC costs. If RC costs go up then we have to implement RC pools. If RC pools are implemented then all the RCs that whales and orcas were wasting will start being used, increasing RC costs even more exponentially via sudden massive hyperinflation.
Once this happens one of two things will occur:
- Users will stop using Hive because it is too expensive.
- Users value Hive enough to pay the cost, raising the value of RCs and Hive.
- This allows nodes to afford bigger blocksize limit,
or as @blocktrades has teased: 1 block per second.
- This allows nodes to afford bigger blocksize limit,
The #1 option will not happen until Hive price is bubbled, so we are nowhere near that moment. I think Hive would have to be trading above $10 a coin at that point before users were no longer willing to buy more in order to compete for RCs.
Also obviously this happenstance will incentivize the entire network to find more scaling solutions. I believe you were remiss not to include the dapp that sucks down the most bandwidth by a huge margin:
Splinterlands
Any casual check of the blockchain at any moment will show that Splinterlands custom JSONs are on every single block in huge numbers. It's very obvious that this can not continue forever. If we had other dapps like Splinterlands on chain the blocks would already be fully tapped out. At some point or another Splinterlands will be way too expensive to play directly on-chain and they'll have to move the bulk of these transactions to second layer or simply move them to the private server where they will not be transparent.
Conclusion
If I'm being honest here I don't think I've "schooled" @taskmaster4450 with this comment in the least bit. I wouldn't be surprised to find out that he learned nothing from what I've written here. When it comes to Hive we both know quite a bit.
I am also not trying to invalidate what @taskmaster4450 has written. I view his method of explanation (in general) as a gateway to the masses. My approach is littered with technicalities that even the people who follow me and have been on Hive for years struggle with what I write sometimes. I'll talk about servers and nodes and API and code snippets and javascript, mySQL, Chrome extensions, node.js, blah blah blah blah blah. Bro, what are you talking about? Stop talking nonsense.
So what's more valuable? Being right or being understood? Does being technically correct matter when the end-message is the same and nobody understands what you're talking about?
To quote myself:
Never done this silly "open letter" thing before.
Let's see if I get the gist of it.
"The gist of it..."
Exactly. Sometimes that's the most important thing to convey. The substance of these messages are the same: real-estate on the main chain is super valuable, and the network isn't going to realize that until we run out, which could happen at any time.
On that note have you heard that @blocktrades and @theycallmedan may be airdropping tokens based on Hive stake? Two possible airdrops and an RC shortage... hm... things could get pretty crazy around here if you ask me.
Posted Using LeoFinance Beta
We'll jump on Koinos chain.
Problem solved, hehe.
Posted Using LeoFinance Beta
ha! nice.
Koinos is going to cater to this community quite a bit though.
The vast majority of users with Koin are Hive users.
And they have already stated development for platform-independent libraries.
Really appreciate this open letter thing.
🙏😎🥓👍
Highly Regard you as a witness bruv.
👏🙏✍️🔑
Thanks friend.
I have a witness team.
We are @hextech.
I haven't been hyping up our witness much though because I haven't been programming.
Start hyping it up boy.
Well I was keepin it secret for reasons but...
https://giphy.com/gifs/warnerarchive-sci-fi-xTiTnhzabxwpofKIeI
I am schooled.
Posted Using LeoFinance Beta
Or anywhere in between.
Don't complicate my argument with an entire market! Dick!
Supply and demand! Get out of here with that!
:D
🤣🤣🤣🤣🤣🤣🤣
!ENGAGE 25
!tan
ENGAGEtokens.I would 100% agree, but I think hive can at some "danger points" easy adopt. Our government settings allow us way easier to fork/update than most other cryptos.
If the price is super high for transactions, systems can be updated. Also, I think with time computer power goes cheaper and cheaper.
I don't see a huge scaling issue there. More on other points:)
Thanks for the good Explanation :)
Posted Using LeoFinance Beta
Hmm very interesting.
Posted Using LeoFinance Beta
Is it? I'm in a weird manic state of mind right now doin all kinds of crazy stuff.
Posted Using LeoFinance Beta
Something I have long wondered... can RCs every be "re-valued"? For example, if there was a ton of activity like you talked about above, one that lead to RC shortages etc, could developers simply go back in the code and revalue the RC to HIVE peg that they originally coded? Is that a possibility?
Good question. I'd like to see a detailed explanation of how the RC system works and how it is connected to the physical resources that is prices.
One fails to see how there will be enough demand, the other fails to see how there will be enough supply. I guess that is why there is a market. ;)
True words. Honestly though it's just downright hilarious... like... how many times does this network have to prove we have HARDCORE support at 10 cents before the mob stops being bearish at 10 cents? Infinite?
Who is this mob, and can you quantify their impact?
I've seen plenty of finger pointing, but little evidence.
The ratio of people who are bullish/bearish on Hive as it pertains directly to price is vast. At $1 we are an unstoppable superpower and at 10 cents we are going to zero. I've seen it happen like half a dozen times already.
No one is going to be able to break away from then mentality until our 10 cent support gets boosted up a bit from usage and fundamental gains over long periods of time.
Right, but somebody is selling, who they are, and the size of their impact, seem like something I would like to know for sure.
We can stop voting rewards to them.
We would be at 5usd right now, but folks keep selling us down the river.
I prefer to maintain a free market. Punishing someone for dumping coins is kinda like flagging Bitcoin that was used for drugs. It undermines the free-market, fungibility, and entire point of the system. Whoever is dumping: it's not their fault that none of us had any liquid coins when Hive was trading x10 at $1. Accounts that sell low don't need to be punished any more than they already are by the market.
We could always guess who's dumping by who sends what to the exchanges in bulk while the price is tanking... but again you'd just be guessing unless you got help from the exchange to violate user privacy.
We could compare lifetime income to current holdings and know, pretty much, who is cashing out.
If stopping the flow of blood from our open wounds is out of bounds, this is gonna be harder than it has to be, imo.
I haven't heard this? Where was that posted?
I think you and taskmaster are saying the same thing, just using a different language to get there. I can only understand the technical part up to a point so the slightly less technical version works for me but you explain things well in your own style. Having read both yours and Taskmasters post I actually learned more anyway so kudos to you both!
Posted Using LeoFinance Beta
no idea for the blocktrades talking about airdrop, theycallmedan was talking about it here https://peakd.com/hive-181335/@theycallmedan/hfjjxupc it is a long video but it is an interesting conversation.
Posted Using LeoFinance Beta
Ah ok, thanks a lot for the link! I'll check out the video :)
Posted Using LeoFinance Beta
TLDR LOL
Nah I getcha. RC stuff is complicated but if we scale more, things like Splinterlands will definitely need to be offloaded to something else, maybe it could be the start of a layer-2 DPOS built on-top of layer 1? That would be cool.
Why did you hide that gem at the very end? I skipped past it until I saw it was in Wombat's comment lol. I will have to look at their mention of an airdrop. Airdrops are always nice!
All the Splinterlands transactions are just text stored to the chain. A Hive node does nothing else about them but store them.
Why is everyone suddenly so damn worried about "blockchain bloat now"? This is patently ridiculous. Isn't activity PRECISELY what we want here? The question is that if transactions are so damn expensive, then why are they not priced properly? Hive Power gives you a right to fill a certain proportion of a block for life. All the transaction costs are being paid by speculators or ONLY new users who buy the token. Without growth the system is FUCKED because no growth means NO MONEY. But growth also means more usage and more space on blocks used.
Is someone trying to bullshit someone or should Hive consider introducing transactions fees?
Yeah I’m not too worried about any of it because I’m not on the technical part of it. Just thinking out loud about the additional layer of DPOS linked to the main chain would be cool but I haven’t the slightest clue if it’s possible or pointless. I just keep using the chain and worry about things I know.
RCs are transaction fees. They just have a zero value at the moment. The governance resource has to be separated from the bandwidth resource for maximum agility. The system we have is good, in theory. We'll be testing that in the real world soon enough. We'll have to pivot if it sucks.
Posted Using LeoFinance Beta
You only have to pay for your RCs once when you purchase your HP. They replenish for free forever when you've paid for your HP.
well if hive would be 10$ and you would need 50 hive to use it at a "normal" level that would maybe be a problem :D
but that would probably be a lot more transactions then we had in the best times of the blockchain history? so that would probably be a good problem to have. and i assume that witnesses would not complain that much about ram usage if the price was 10$
didn't see blocktrades talking about airdrop. heard dan talking about it for 3speak.
This is why RC delegations are essential for future growth. Provide the ability to transact for a lower start up fee than 50 Hive x price.
Posted Using LeoFinance Beta
So many technicians, I haven't really thought about HIVE being too expensive to function. Maybe add some fees to the network? Would that not be functional?
I also believe expanding the top 20 to maybe top 50 would that not be functional?
might be more or less simpler than we think.
It is not about being right or wrong... it is, do you understand.
Posted Using LeoFinance Beta
RCs are this network's fees. We have so little demand (or such good scaling) that those fees currently round to zero. The system we have is good.
We can't start paying more witnesses until we have more stability and witnesses are getting paid an obscene amount. Instead of a 10 cent floor it would need to be something like $1 before we can talk about that. Again, paying more witnesses increases our decentralization, thereby making our scaling worse.
Network fees for smart functions instead of SMTs.
20-50 is a great suggestion
Posted using telepathy
The total number of transactions is usually between 20 and 40. Then there's about an equal number of virtual operations that the blockchain does on its own without being initiated by any user. What percentage of the 65 kilobytes are we talking about?
Look at the blocks: https://hiveblockexplorer.com/
Aggroed mentioned that running a Hive node costs about a couple of hundred dollars a year. An RPC node is a bit more expensive. That was around last Christmas. That is not too expensive. A top 20 witness still makes more than $10,000 a year even at this kind of low HIVE prices.
You are part of a witness team. You should know how much running a witness really costs. How much?
I find this to be the most valid question.
What's the cost in ratio to amount being paid to block producer? Is so significant that block producer cannot easily afford to increase their capacity if need be?
Posted Using LeoFinance Beta
Running witness node only costs a few hundred dollars PER YEAR at present. An RPC node (a node that DApps make remote procedure calls to when they interact with the chain) costs a bit more.
A block is signed once every three seconds or 20 times a minute on Hive. A top 20 witness signs a block thus roughly once a minute. In a year, there are 60x24x365 = 525,600 minutes. The block reward is 0.261 HIVE at present. Thus a top 20 witness gets about 137,200 HIVE as block rewards. The price of HIVE is 0.115 USD at present. Thus a top 20 witness makes about 16,000 USD a year at present. Relative to the cost of running a witness server, witness rewards are plentiful.
Hive should be prepared for onboarding millions of users. But after all these years I would say that some attention to should be paid to how all those users should be onboarded.
Posted Using LeoFinance Beta
Pennies from heaven? Yes, please.
Very well explained, but far more important than all of this technical jargon, have you muted JRCornel?
Why do you ask? He seems to be unfollowed on Peakd for some reason.
Posted Using LeoFinance Beta
It's just his comments were hidden - OH no HANG ON - It's 'cos I muted him, I'd forgotten! (Too many posts simply copying what other people had said with too high pay-outs just annoyed me).
Green Monster... :) JK Bro
Posted Using LeoFinance Beta
Not really, I'm delighted to see people get high pay-outs when they deserve it!
When they don't I cringe because that's what almost drove me away from Hive!
I hope we get muting on LEO asap there's several people on trending whose posts I'd rather not see!
you could downvote
Posted Using LeoFinance Beta
That just runs the risk of retaliation! I'd rather not. I'd really rather convince people to not vote for repetitive material in the first place.
In fairness to the guy it's not all like that, but a lot of posts are very milky, and he knows it.
But what if he sees your comment, won't he retaliate with a downvote?
Why would he do that? He's not an unreasonable individual - he's not the type to DV me for stating an opinion!
I don't want to DV him as it's not every post he does, I just need to not see his almost copy-paste and very brief posts earning $20 a pop, it's a joke, people voting stake. It's not really his problem, it's every one else upvoting him - probably a result of AV on Hive carrying over to leo, or just people voting stake.
NB He's also obviously a pretty intelligent guy - I can guarantee you 100% that he knows EXACTLY what I'm talking about and agrees with me. There is no way he actually thinks those types of post are worth the money either!
I mean I cringe at my own pay-outs sometimes, he's just another tier up, same thing really.
When thinking about this kind of thing
I sure hope the air-drop is HP only. https://hive.blog/hive-167922/@jrcornel/what-is-going-to-drive-demand-for-hive
I’ve thought about the problem of Splinterlands blowing up and running out of bandwidth. What is their best solution? Fork the chain and let the community run separate witnesses in exchange for DEC?
Posted Using LeoFinance Beta
To be honest... the solution is only storage decentralization to be included on the blockchain code. I know... a big F problem to solve but in the long term the only valuable one.
There are already many blockchains techs trying to solve that "problem". Even if not everyone can participate with the same kind of gear, finding an alternative so that small fish can help the blockchain on one of the most community based and extensive blockchains in the entire world, is a MUST!
I have fortunately lots of skills and hardware that I could just run a couple of nodes if I wished so. I even had done that in the past while on STEEM, on EOS, and then I stopped. Because I saw some promising path being planned for distributed DB storage into the blockchain part. I don't care much about the API yet query capacity yet... that one can be centralized easily because it can be "rebuilt".
But the core function of a blockchain, if centralized, by itself is not scalable. Bitcoin fails at that... or by any means any sort of PoW or PoS that relies on database consensus for forking reasons.
Yes there are other techs that solve these problems, but they lack other stuff... so its always a balance. The point is, on an already proven for years code, the next step needs to take place if we wish to compete with things like Google, Twitter, etc... And that SHOULD be what the community should be aiming for!
We are already out of the charts in terms of historical breakthroughs... so why struggle with just "surviving?"...
Posted Using LeoFinance Beta
This post helped me to understand RCs much better than I did before, you earned a follow for that!
Could you make a post on the current data in blocks? Which apps fill how much of the block, how much more they can handle, etc...
Posted Using LeoFinance Beta