In section 4.2, the Chainlink whitepaper identifies the fact that on-chain aggregation of oracle responses on a permissionless blockchain (eg Ethereum) will be problematic from a cost perspective because each action performed in the commit/reveal function will accrue network usage fees (eg Ethereum fees).
The Chainlink main net will launch with on chain aggregation (off chain is slated for the medium/long term) which seems to suggest that high volume use cases will not be appropriate for some time after main net launch.
Is this a bottleneck for Chainlink network adoption, or is on chain aggregation likely to function effectively enough for the intervening period before off chain aggregation is implemented?
In section 4.2...
Wyatt Torres
Ian Lee
Good question
Reading over that section (4.2) it sounds more that the issue will be with use cases that require high levels of security rather than 'high volume use cases'
The bit that's expensive on-chain is aggregating the node responses. The fewer nodes you specify, the lower the fees (because there's fewer to aggregate). It's not a question of high volume, it's a question of security: more nodes providing the data = more security, more difficult to tamper, more reliable, etc. but it's also more expensive the more nodes you require
I might be wrong but that's how I see it
Christopher Cox
For us to make it it would be more than enough if swift or some other big guy like AXA just started testing it publically. I stand together a la iota and wv would ez drive the price up to 100+USD
>le banker transfer linkchain coin
Grayson Anderson
If this is correct, it's not such a big issue.
The amount of security/tamper proof-ness you require is likely to be proportional to the value of the smart contract. If your smart contract is for $1Bn of bonds then you're going to want as many nodes fetching the data as possible because that way it's more secure - but it's also more expensive. If your smart contract is just some $10 deal then it's probably sufficient to just use 1 or 2 data sources (and so 1 or 2 on-chain fees) because the use case doesn't demand a high level of security.
Sure, it would be cheaper to aggregate responses off-chain as per the long term strategy but settling a contract on-chain in the short term way is still likely to be cheaper than hiring a bunch of lawyers etc
Jaxson Davis
The section you posted as an image is explaining that when all the oracle nodes perform their commit/reveal function (which is where the individual nodes sign their answer, and when a certain number of answers have been signed they all "show" what their answer is, to prevent nodes just copying other nodes' answers) using on chain aggregation, all of those commit/reveal functions need to be performed on the Ethereum blockchain, and accrue network usage fees as a result. It's definitely a bottleneck in regards to high volume applications (where lots of commit/reveals are required) rather than security.
William Moore
ah right, sorry
I didn't read your OP closely and interpreted you to be bringing up a cost issue. Reading it again you're saying aggregating these responses on chain is a bandwidth bottleneck - i.e. the ethereum network is too slow to deal with all these commit/reveal requests - is that correct?
My response to that would be that if chainlink is so wildly popular that it creates a bottleneck in the ethereum network you'll have bigger things to worry about - like what color lambo you want lol
Austin Baker
No I'm saying that it is a cost issue. If every single commit/reveal has to come with an Ethereum network fee (cost of gas) then there is a strong disincentive from using contracts that need a high number of on chain aggregation responses.
Off chain aggregation solves this because all of the commit reveals happen, and then the "final" answer is sent to the smart contract and it only accrues a single fee, but that won't be the mechanism used at main net launch, and is not slated until the medium/long term.
Liam Adams
>If every single commit/reveal has to come with an Ethereum network fee (cost of gas) then there is a strong disincentive from using contracts that need a high number of on chain aggregation responses.
This is true. But the contracts that need a high number of chain aggregation responses are going to be the contracts that are the most valuable - e.g. $1Bn of bonds. And since these contracts are so valuable the people creating them are likely to accept high network fees (at least until chainlink can do this off-chain).
It goes back to what I said here >The amount of security/tamper proof-ness you require is likely to be proportional to the value of the smart contract.
Sure, it will be a disincentive for low-value contracts to use many data sources. But there's a trade-off between cost and number of data sources. High value contracts will accept the high network fee costs because it allows them to use more data sources i.e. it makes the triggering of the contract that much more secure
Hudson Davis
Basically: The more data sources, the more security. But data sources cost network transaction fees (until chainlink can settle this off chain). Low value contracts can opt to use fewer data sources (and so save network transaction fees) but the contracts will be less secure. High value contracts can opt to pay for more data sources and thus more security.
It's still orders of magnitude less expensive than hiring lawyers and people to settle these contracts via the current system.
Aiden Myers
congratulations lads on the non-brainlet LINK discussion
gonna make it
Jordan Adams
This. Was a refreshing read.
Nathan Lewis
You are absolutely right mate. The last tweet of Sergey Nazarov was a mention to gastoken.io a project that adress this issue and might try to solve it.
>twitter.com
>gastoken.io
Hudson Roberts
Brainlet here, do we really need more than maybe 1 or 2 trusted sources for smart contracts to execute?
Can't companies just enter agreements with each other that gives them a way to back out in case of obvious error?
Example being you are trading and using a smart contract that needs a share price. I think the link wp used Intel as an example. So let's say you use the official nasdaq site and it fetches the quote and intc is $30. It triggers some trade automatically. But actually the true price is 35. Okay, that happens like 1/10,000 times, a simple error of some kind that you really don't see often. The trade is reversed. The most common thing would just be having accurate data and it goes through normally.
How often does weather.com give you the wrong temperature?
How often does the Wall St Journal quote the wrong interest rate? Not often enough to wait a year for development.
Prove me wrong.
Levi Edwards
real discussion? No pee pee poo poo? Color me impressed! We did it Reddit!
Adam Gray
>do we really need more than maybe 1 or 2 trusted sources for smart contracts to execute?
>just enter agreements with each other that gives them a way to back out in case of obvious error?
These two questions are related. You're missing the whole point of chainlink: by using multiple independent data sources you decrease the likelihood of 'obvious error'. If you only used 1 or 2 trusted sources then there it would be more likely for there to be cases of obvious error.
The second point:
> agreements with each other that gives them a way to back out in case of obvious error?
The whole point of smart contracts is to avoid this situation. In the current situation companies always try to back out when deals they've agreed to don't go their way. Often, even when one party is legally owed money by the other side it's not worth going to court because it's so expensive. There are tons of disputes that get settled out of court for less because it's not worth the cost of years of litigation.
The whole point of smart contracts is they execute AS WRITTEN. Once on the blockchain neither side can back out unless BOTH sides agree. If you had a situation where one side can 'back out in case of obvious error' it would defeat the whole purpose of smart contracts because if one party didn't like the outcome they could troll the other party by claiming it's an obvious error.
'Obvious' is a term open to interpretation. What might seem like common sense to you is the kind of point lawyers waste days months and years squabbling over in court.
Ethan Fisher
The Problem arises when one party is not willing to settle the error.
Let's say you buy 50t of steel from the chinese and state in your contract it has to be delivered by July 1st. You two agree to pay half in advance and the other half when the contract is fulfilled. The steel arrives at July 2nd. The contract was obviously broken, but the settlement will include multiple lawyers working multiple weeks on both sides. If you would have set up a smart contract this would not be necessary.
Justin Robinson
in this specific case how would the smart contract settle? You clearly don't want a situation where the party receives the steel on July 2nd and pays nothing for it just because it arrived a day late.
I guess the contract wouldn't execute by midnight on July 1st and they'd have to negotiate a new contract...
There's also the issue of how you could digitize receipt of steel even with a product like chainlink
Juan Hughes
You two have described the If & Then clauses, but I guess you could also include or negotiate an Else clause in which you receive a % discount for every day with deliver delay inside the smart contract
Aaron Smith
I imagine using XYO in some capacity would fix the tracking issue.
Brody Scott
yeah that would be a pretty simple solution and easy to program
Robert Smith
Keep the shilling in check Linky
Or I'll bog dump my Fud bags
Going to sleep right now. I'll check in tomorrow
Ayden Campbell
The world today is far away from being able to decentralize sources for physical phenomena like delivered items, unless you can somehow guarantee that nodes in many different locations are somehow watching the delivery location of the steel AND are smart enough to tell that it is the correct amount, material,quality, etc. The closest we can get to the physical world as of now would be weather information. Smart contracts with chainlink are largely limited to APIs for now until someone can solve the “just find the sensors and disable them bro” problem
Aaron Perez
Also machine learning is still shit right now
Tyler Hall
Currently when you order something from aliexpress you can track the delivery status every time it cross a checkpoint. I think it would function like this but sending every data to the blockchain instead of sending it to a database.
>pic related
Camden Ramirez
Also for example DHL could open their API and sell its information to node operators.
Jayden Jenkins
Ryder Moore
That might work if you didn’t care about what was being delivered or the condition of the package contents. You could be getting a pile of rocks instead of the computer you ordered the smart contract would self execute. Even then you don’t truly know where the package is because how many carrier APIs are tracking this item?
Austin Kelly
It reminds me of the old discussions which we used to have. The ones which actually made us pool the money to get in on the ICO.
Colton Allen
Yes they can but at the end of the day the incompetent delivery drivers can press that “delivered” button before they even leave the warehouse. That is not something anyone who is smart with their money would want to be allowed in a smart contract.
Grayson Sanders
Someone copypasted your post on Slack and asked Thomas. He replied
>pic related
Blake Walker
Thanks for sharing this.
I have four questions:
1. Who will pay these costs, is it the user who requests data or is the dataprovider? Or is this something that can change depending on which smartcontract you use?
2. Right now it takes several minutes to send something via the ethereum network. Isn't this too slow if you want say the BTC/USD price? Due to high volatility the price can change several % between request of a value and the receiving of the result.
3. Is there any idea how they can prevent reputation farming? Let's say I make a node providing the air temperature in some god forsaken hellhole. I make non-stop requests for this temperature, and even though the network is decentralized I'm a 100% sure I'll get a response from my own node. Hence, I could boost my node's reputation to unseen limits.
4. Why stay on the ethereum network? Wouldn't a proof of stake based own network be more secure/faster? Long term ofcourse, in short term I understand they stay on the ethereum network. But they don't mention going their own way in the whitepaper and this always suprised me.
Henry Myers
>Delivery nigger is too lazy and tries to press button he delivered
>package is not and API, GPS is proof
>niggers delivery car gets send back to warehouse automatically
>nigger gets out of delivery car „ayo hol up why me back in warehouse nigguh?“ he asks
>supervisor gets the security and they drag the black man onto the street
>white man gets behind the wheel and now we get the second delivery attempt
>only if customer signed that he received the real goods the smart contract will pay out the delivery service so it is delivery companys incentive to deliver to get the money
>turns out the chinks only send rocks instead of cheap steel
>customer rejects the package and does not sign that he got the goods
>half the payment money is automatically delivered back into customers account (funds are frozen til confirmation of delivery)
>chink account is flagged as not trustworthy and chink loses money on this scam, chink has no incentive to scam
Adrian Hill
>t. Doesn’t know how to use a smart contract
Jacob Jackson
t. Low IQ fudder
Charles White
>writes worse than a nigger
>still blames niggers
Christian Gutierrez
offended Nigger detected, go play some B ball homie.
Lincoln Jackson
Your fud is retarded and only succeeds in drawing more attention to the project.
Just like in real life, you are a failure.
Blake Murphy
Axa made there own oracle according to Jonny
Jacob Campbell
How about this scenario
>delivery arrives at nigger mcniggs doorstep
>nigger beats up the whitey delivery guy, takes legit package and refuses to sign the contract
>nigger is fully refunded AND has free package
>there isnt jack shit the company can do to get the money back
Angel Miller
Chainlink = /pol SCAM group
>be conspiratard polack
>spend years there
>learn all the merchant tricks
>take over /biz with LINK
>scam your fellow people
The absolute state.
A new generation of bagholders is created with every new thread. These bagholders will become scammers themselves. Gotta give it to polacks, they learned from the best merchants.
Levi Rivera
>cops shoot nigger
>money is refunded because contract wasn‘t signed when it was supposed to
Jack Green
It's good to see some actual discussion for once on this board. Is it true that anyone could create an oracle easily? Chainlink has some competitors now with Qubic and i like both projects desu
Nicholas Scott
What if cops dont shoot the nigger because cuckstate took their gus away
Kayden Kelly
>Is it true that anyone could create an oracle easily?
it's difficult to do what LINK is doing, which is building a secure block-agnostic oracle that will handle high-value contracts and incorporate a reputation system.
building a bog-standard oracle which doesnt do those things is an undergraduate-level project, so yes, pretty easy if that's all you want to do.
Noah Gonzalez
Then nigger is arrested and moved to the interracial breeding grounds.
Lincoln Kelly
Yes. Look at my post above.
Axa is the source of hope for many linkies but they have made their own oracle
Tyler Cook
If thats such a difficult project to do, why are there only two guys actively working on it, they are not the most accredited devs either. We need some super minds working on this
Grayson Martin
I love these threads but I’m too much of a brainlet to understand
Jayden Kelly
Sweden?
Joshua Moore
>If thats such a difficult project to do, why are there only two guys actively working on it, they are not the most accredited devs either. We need some super minds working on this
there's more than two guys working on LINK. ari juels alone probably has a whole cohort of PhD's working on it over at Cornell. LINK is way bigger than what most people realize.
Cooper Gutierrez
same here brother
>7.5k linklet
Jonathan Flores
Decent smartcontracts are insanely difficult on a Turing complete platform. Most (99%) of all smart contracts are either baby's first smart contract or buggy crap that's not suitable for anything. Probably the chainlink team is slowly building a database of ready to use smart contracts so their main net can be used from day one. If not, adoption could be a long way out.
Anyone please refute my fud.
Austin Davis
They are simply technical advisors
>get a call from Sergay to check over the code once a month
>yeah sure whatever fatty
>run over it in a minute
>say everything looks good
>get back to fapping
Camden Davis
>They are simply technical advisors
no-one works for free, but i only know as much as you do, and that ari juels has a team working on it over at Cornell, LINK has hired and is still hiring new talent, and that if microsoft has them working on their own projects there must be a staff crew behind the scenes.
or maybe not.
Cameron Morris
>10k linklet here
Feelsokayman. Each week I throw in $100 from my wagecuck paycheck
Tyler Richardson
Amazing thread, thanks op for this
Josiah Campbell
How many linkies to get out of linklet rank
Eli Foster
25k
Blake Howard
Thomas replied again.
Answer to question 1:
The person requesting the data will pay for the costs associated of all nodes submitting answers to the blockchain. Currently, we only have one config option for determining if the node should take the request or not, but that may change in the future. The node operator should set their `MINIMUM_CONTRACT_PAYMENT` amount to factor in for all of their costs to write to the blockchain.
Answer to question 2:
There are additional factors at play here than just volatility. You could also have a difference in price between multiple exchanges and pairings. This is something that I think would fall on the contract creator to consider when choosing their data sources, rather than something that we (or any oracle service) would need to solve. The Chainlink node could provide millisecond-level updates if the API and blockchain would support it.
(Cont)
Logan Lee
Answer to question 3:
Allowing for multiple reputation providers is one way to prevent reputation farming. For example, if you've given yourself a bunch of requests on one reputation provider, that doesn't carry over to another provider which may be highly regarded for valuable smart contracts. However, I can see the concern of someone giving themselves requests that only they can fulfill on the targeted provider. If the requests for data come entirely from on-chain, this would be fairly easy to detect by an on-chain security service, or anyone wishing to validate the history of a node's requests. If the requests are coming from off-chain (like if the reputation provider has an oracle marketplace), it would be up to the reputation provider to take action if there is something malicious going on. However, there will be some legitimate use-cases for using a single node (or multiple from the same operator if they provide some sort of redundancy) multiple times for specific tasks that aren't available to everyone. Hopefully that at least addresses the concern. Concrete details of how reputation will be implemented on-chain will be available later, likely around when a node's reputation is measured to fulfill requests.
Answer to question 4:
Simply put, Ethereum is the leading smart contract platform. Creating our own network/blockchain would add a *lot* of development overhead that would be distracting from the main goal of the project.
Oliver Brown
Oh look EOS solves all this.
Jacob Phillips
>EOS
Jack Robinson
Thanks a lot (seriously user).
Great answers. Thomas is either really intelligent or a great anti-fudder (or both).
Landon Lopez
You are welcome user. We are all, unironically, in this together
Daniel Barnes
Are you in the discord?
Chase Cruz
Nope, but anyway I'm not interested in discord groups. Only community I'm in is Slack
Nicholas Parker
What am i looking at here?
Justin Martinez
Okay, otherwise you'd be a great addition. Thanks again and see you in the next decent link thread.
Josiah Lopez
You can't even name 1 thing chainlink can that EOS can't.
Colton Ramirez
Chainlink is blockchain agnostic
Jose Reyes
Gfy and go buy EOS then.
Gavin Hernandez
kek, you literally can't name 1 thing can you?
James Cox
Not have massive bugs and glaring security flaws days before release. Not allowing your tokens to be stolen based on a vote by nodes. Not being foced to give up your private key.
Gabriel Thompson
EOS isn't connected to Swift as Link is, nor does it have the ability to automate the messaging process Swift supports. EOS may make it, but end of day is not going to be as necessary or as successful as Link.
Lucas Roberts
Edited to add, you should invest in EOS though as Link may be too complex for you to comprehend.
Kevin Turner
A+ questions, mate. High energy
John Collins
When you guys say 10k or 25k are you referring to individual links owned or how much fiat you've put into it.
Pls don't laugh
Jack White
Typically they're talking about number of link tokens.
Though one "Link token" is really 10^18 Link tokens. If you read the smart contract on etherscan.io for the Link token, you'll notice the total supply is 10^27, which is a billion "link tokens" that are each divisible 10^18 times.
James Brown
I'm not autistic enough for this.
Seems like Link is about 25 cent. It doesn't seem like getting 10K would be much of a investment at all. Are people expecting a moon or linear progression.
Robert Sanders
I literally might drive from the southwest of America to Cornell on a cross country road trip (with my AK-47) and personally ask Ari if we will make it
Julian Jenkins
we're poor NEETs. 10k link is an enormous amount
Luis Robinson
The number of links you actually own. And I'm taking a massive shit as I type this.
Leo Cook
Not everyone got in at 25 cents though. A lot of these guys paid $5k-$10k for their 10k Link tokens.
And quite a few of us bought in between 13 cents and 30 cents.
Matthew Brown
I see. Sorry I wasn't meaning to come off as haughty I was just curious about this LINK stuff that gets spammed everywhere.
Andrew Davis
There's no apologizing on Jow Forums.
Elijah Miller
There are black people that own link who didn't pay over 1$ for Link because they don't have the money but are smarter than those who bought over a dollar because they dont have the money to pay over a dollar.
So remember that link marines who paid a dollar or more for Link. There are black link marines who own more Link for less than you bought it at
Matthew Kelly
Many of us are unironically expecting Link to hit $1k eventually. The EOY thing is a meme, but depending in adoption and a million other unknowns, it could actually hit those prices. So 10k linkies would make you pretty wealthy if you managed to hold on through the mission to mars
Connor White
What the heck do you mean you didn't mean to come off as naughty????
all you asked was does 25k mean 25,000 dollars invested or 25k Link
Jacob Murphy
Not have a pedophile CEO.
check mate
Andrew White
>tfw i always stack my buy now.
I learned a hard lesson when i got in a .89
Ryan Watson
Don't forget sharpies
Aaron Carter
I have such a nice ass
Robert Hernandez
Can’t you just led nodes aggregate fees over the course of a day/week and pay out the total all in one transaction when the node operator requests it?
Charles Torres
And then all the clever fellows in Chainlink have sold at the SIBOS because they understood that nothing will be out for ages.
Aaron Wood
it depends on what you mean by clever. some people dont know how to ride pumps. ITs safer just to spend 10k and forget it.