Dr. Leemon Baird: How Hashgraph Works - A Simple Explanation w/ Pictures

NOTE:
At 43:51 Dr. Leemon Baird says:
'Now C2 can't see it - there is no line from C2 to the black event."
- This was a mistake (C2 can see [Black Event] through Alice); however, it doesn't matter because C2 is not famous.
Timestamps for this video
0:00 Intro
0:33 What is a hashgraph?
4:10 Why use a gossip protocol and how does it work?
6:44 What does an event include?
10:11 An example of a hashgraph
12:33 Rounds of voting without votes
15:47 What are witness events?
18:18 What are famous witnesses?
23:32 Counting votes are based on strongly seeing events
31:47 How to determine if a witness isn't famous
36:15 How are normal rounds and coin rounds different?
39:55 The magic is in the math
41:14 Swirlds SDK download www.swirlds.com/download/
42:36 Sorting all events
51:14 An example of a real hashgraph demo and closing
Join our community to learn more:
Website: hedera.com
Twitter: / hashgraph
Telegram: t.me/hederahashgraph
Discord: hashgraph.com/discord
#Hashgraph #LeemonBaird #VirtualVoting

[@benyaminewanganyahu]

How do I play this on bass guitar?

[@enki-not-enlil]

Skip the hashes and look at the timestamps! You only need the hashes for playing in the band! :P

[@code_kanga5390]

benyaminewanganyahu best comment ever

[@SamsonZoomBespoke]

very funny|! But invest and you will find harmony and har money

[@mistersmith8962]

Tune that bass to A,B,C,D and hash it out ....its by a band called The Speed of Gossip

[@U2Berror]

why is everyone complaining the simple video of this was 52 min! I would love to see an 18 hour video on this!!

[@vivaparenzo]

You can hear it is a great professor talking, subject is so natural to him, he must concentrate hard to put knowledge in simple words.

[@HederaHashgraph]

Hi vivaparenzo ! What a nice comment! Thank you. We agree and are extremely proud of Leemon - not only for his expertise and knowledge - but also for his ability to break down difficult and complex topics into laymen's terms. If you want to stay up to date, check us out on Telegram: t.me/hashgraph Newsletter: hashgraph.com Forum: hashgraphboard.com

[@juri8723]

@@HederaHashgraph „into laymen‘s terms“ almost sounds like „into leemon‘s terms“ lol

[@labado5969]

Awesome 👏 Next step is to make a extremely simple and short animated move in 1 min for all people to understand ✅

[@ryblinkaway]

Thank you for saying that La Ba Do!!!...oh my god!...can this be explained in layman's term? not everyone is as genius as Mr Baird!...

[@danieljonce]

People calling for Hashgraph to "open source the tech" and upset that they "have a patent" have disregarded the possibility of a defensive patent and are confusing different types of intellectual property. There is a difference between U.S. copyright law and U.S. patent law. Copyright law protects specific works automatically immediately after the work is created. There is no action one must take to claim protection. You can "register" a work but that only reinforces the protections that already exist. Open source licenses were created to allow creators of works to specifically allow public use of such automatically protected works. A work could be a picture, a song, a book, or a block of code. Ideas are not protected automatically like works are, rather you need a patent to protect an idea. This could be protection for the purpose of personal gain but it can also be protection with public interest in mind. In fact a great way to protect an idea in the public space would be to first successfully patent the idea and then specifically release it into public via a free public license. If you don't first have the patent secured before you release an idea to the public someone else could conceivably patent the idea and then they would control over the use of the idea. I don't know the motivation for the Hashgraph patent but I know it's too early to judge it. The indicator of the motivation will be how Swirlds decides to license the protected idea to the public.

[@hugopristauz3620]

I looked for an explanation how hash graph (really) works - now I found it! Thanks!

[@NoticerOfficial]

Flash forward, May 2021, Hedera Hashgraph reaches 1 billion transactions in 18 months, alongside Bitcoin’s milestone 700k transactions in 12 years

[@tamarab6680]

Thanks Dr. Baird for this great explanation! I read the WhitePaper and this makes so many of the concepts much clearer.

[@baobabkoodaa]

All of the important stuff was "technical detail, just trust us, it works". I can't believe this video is 52 minutes.

[@Supremax67]

Only found out about Hashgraph yesterday. I have to say, this stuff is nice. Security and speed.

[@victorlenus]

Really nice and simple. Don't mind the time, it is worth.

[@sandstorm973]

Thanks for breaking this down. I'm interested in learning more about how to develop applications using this technology. Are there docs available? I've downloaded the SDK.

[@samuel28858]

A simple explanation of Hashgraph, hmm i wonder how long it is, oh only 52 minutes.

[@MatthewMarcus]

Leemon did drone on a bit and repeated himself several times to ensure the point was well received. Also, it seemed like he had a cold or tickle in his throat which added some time to the video. Overall though, pretty well explained.

[@benyaminewanganyahu]

Is it not obvious that, by definition, anything complex, explained simply, must take longer to explain than when explaining it in an elegant, complex way? It's refreshing to have such a thorough explanation in a video format by the author. Absolutely no complaints about the time length here.

[@hgfuhgvg]

I had no trouble understanding, and the nice thing is I didn't get bored for the entire time.

[@damienwins6185]

exactly. A simple explanation will take longer than a complex explanation because you have to stupidify the complex words. Each sophisticated word could take a multiple sentences to explain. lol

[@robertweekes5783]

Don't expect a PhD to give you a simple explanation !! lol

[@DeveshFuse]

Best video on hashgraph you have wide understanding

[@Plowstreet]

Surprising to read comments suggesting that Dr. Leemon Baird doesn't know what he's talking about or just uses this video for the cause of marketing. I think no one of you making these comments have had a look at the already established work of Dr. L. Baird. I'm not a technician myself but I surely am interested in the further development of Hashgraph.

[@bnsgmng]

Very interesting video. I have just one question. At 43:45 Leemon says that C2 can't see the black event. However, C2 receives a gossip packet from Alice's first event after A2. Does this not contain the black event?

[@mruminer]

Very nice. Looking at this it looks to me that there is an error in that C1 would not be in round 2 with A1, B1 and D1 as stated but in some later round as all the famous round 2 events cannot see C1. Am I incorrect? I am not yet going to get into thinking about what that means for consensus timestamp for that C1 event and ordering of the transaction but it does make me wonder about timing attack vectors and a small set of nodes working disconnected from the graph and rejoining or even impact if a node communicates to a peer, drops and that peer in unreliable as well in communicating (guess each node needs to communicate to sufficient peers to statistically remove that issue - i.e. a network effect - imagine that) . The devils in the details of course and this was an example created for demonstration of basic concept. All in all seems promising if timing attack vectors are solidly understood or rendered ineffective for it use cases.

[@hashimer5938]

Love your work!!!!

[@Mikenoronha]

The Hashgraph protocol is what we all would love Blockchain to be. The problem is that it's hard to invest in Hashgraph because, unlike Bitcoin, you can separate the protcol from any tokenization. Bitcoin is the Blockchain. The protocol is also the token. Even if Hashgraph launches an ICO, the protocol is not married to the token. But regardless, we desperately need an open-source version of the Hashgraph protocol before the corporate giants sink their teeth into this. If developers don't move fast, Hashgraph-like protocols will either be the most important stocks to own in 4-5 years or gobbled-up and assimilated into corporate behemoths.

[@Supremax67]

This is where Hedera Hashgraph comes in. Public ledger coming out later this year.

[@ozkanb]

I assume the size of those 'events' never grow because they are being hashed? I'm asking this because each new event's knowledge grows because of all the gossip happened in the past.

[@pr4nk5tr]

So if there are 600 nodes, do 400 have to sign each transaction?

[@samsamyiu]

Of course I understanding everything he was saying, he is right

[@davidicke4451]

This is gonna be huge

[@hard2thinkone]

That is an awesome explanation! Thanks Leemon. One quick question on node join/leave: when a new node decides to join the network, how would it join? Does it make event that jumps straight to the latest round? and when a node decides to leave the network or simply shutdown for a while, how would it resume?

[@hard2thinkone]

Thanks for the reply. I'm actually thinking that in a permissioned network node join/leave could also happen right? I'm looking forward to the addressing in a few weeks.

[@XorAlex]

Have the replies been deleted? Why?

[@hard2thinkone]

Looks like it is. I have no idea why.

[@XorAlex]

I guess they want to give only enough information to make people interested, but do not want to reveal too much, to minimize a chance that someone build a competing system.

[@HederaHashgraph]

replies were lost when we switched from a personal account to a brand account. KZbin's fault, basically.

[@Marsgasbrengen]

Great explanation Leemon, thanks so much! Certainly, because I'm not a mathematician, nor a programmer, but I'm always interested in learning new things. I do however have 2 questions about the system: 1. If I understand correctly any device "creating" and "gossiping" an event is seen as a node, which will also receive all of the information from the other node(s) it has been in contact with. This allows all the events to be gossiped around really fast and very diverse, making it a very fast, safe and spread out (decentralized) system. The thing I'm wondering about is the space needed to record the information by each participating device. So the more the system has been running over time, the more information there is to "gossip" about and to be recorded by each device. The more information there is to be recorded, the more storage space is used, right? So for example if I do a transaction with my small mobile device to a huge server-like device, like for example a server of Amazon. How will my small device be able to record all the "gossip" (like billions of transactions) it will receive back (as per the system) from this huge storage device in the connection the 2 make? And maybe I'm thinking about this totally wrong, then please just tell me so and explain where my thinking flaws. 2. I don't really get how it is determined when a new round starts. Do I get it correctly if I think it's whenever there is a supermajority vote established? After that there will start a new round? Thanks so much for some clarification!

[@Supremax67]

1. Hashgraph is not blockchain, once the consensus state results has been spitted it out, it no longer needs to retain the information. Hashgraph is about ordering the series of events that happens across the network, the moment it becomes apparent that they are no longer events that requires sorting, you have consensus and information is no longer kept. New events of newer time stamps are not considered in the current ongoing method of reaching consensus. Each node would look at an event and compare to other events it received, if there is room to be considered for sorting, than the event is included with all other events (the sorting would keep going until each node agrees to 2/3 that the newer events doesn't require sorting for the ongoing consensus) You would never have the need for an Amazon server to do this. Your mobile device would be paired within the same shard (not really covered in this video) of equivalent performing device, so never would your smartphone be flooded with billions of transactions. Visa only uses approx. 15,000 transactions a second and current smartphone would be close to those performance. 2. Has for a round of when it is created, perhaps the whitepaper could better help with understanding the concept. s3.amazonaws.com/hedera-hashgraph/hh-whitepaper-v1.1-180518.pdf see page 57 of 65. Disclaimer: I am not a developer, however I do recommend to join their dev community discord if you have further questions or the Gossip Guy community discord which also has a strong community support around this technology.

[@Marsgasbrengen]

Thanks so much Supermax, for taking the time to explain. I'm not a programmer/developer of any kind and just try to understand it as well as I can from a laymen's perspective. I really appreciate it!

[@Adolphout]

Is there a difference between real world transactions which can happen between any number of participants and the gossip transactions between voting participants? The system is not “open” in that case.

[@ClawsNGloves]

The true test for this tech is in the public space.

[@HederaHashgraph]

Thanks for the comment! Join us on Telegram, t.me/hashgraph

[@FirstNameLastName-sy4kd]

Aren't the colours at 28:00 the wrong way? Because on the green (=the correct way) there isn't an event on the C node?

[@qazawatzirak8018]

@Hedera I am just curious what will happen when a node permanently leaves the network. How proofs will work in that case ? For example, D leaves permanently and all the proofs in its timeline are lost.

[@Supremax67]

Every node has a copy of the state of the ledger. It doesn't affect the network ability to come to consensus.

[@maxvharper]

How do you linearly sort all the finalized events in a way that is consistent across all nodes? I think I missed that part.

[@sheepsus]

uhmm the virtual voting thing

[@kunalkasodekar8562]

Nice explanation😁

[@hgfuhgvg]

Why do we need ordering in a round? Can we not treat it as a set of all events received in one round? Might make it simpler. Also how does this work with changing number of nodes?

[@XorAlex]

I guess the reason is there may be two double spend (or contradictory) transactions in the same round, that are observed by different participants. Ordering events in a round allows to determine which spend happened first, making the second illegitimate. In normal blockchains it is not necessary, because a miner who produces a block makes sure that there are no contradicting double spend transactions included in that block.

[@c.l.369]

thank u so much 4 ur time! i only understand a lil of it, but i can imagine that this technique can be used for some really good stuff for whole earth! economics in very bright spectrum from elementary things like food or living to stuff like luxusitems and so on, good luck, i hope it gets a nice place where it can work well :)

[@simoneast7]

I also asked this on another video, but I'm still wondering about the gossip protocol. If nodes each pick another random node to broadcast to, wouldn't you likely end up with massive duplication, whereby each node receives many different copies of the same message? And since it appears that nodes broadcast their message multiple times, how do the nodes know when to stop broadcasting?

[@Supremax67]

Nodes will never stop broadcasting, but there is a timing and a logic to their madness. For example, no nodes would gossip 1000 times a second

[@simoneast7]

@@Supremax67 Thanks. Do you know of a document or video that explains this? Perhaps it's part of the gossip protocol, more than the Hashgraph protocol itself?

[@Supremax67]

@@simoneast7 -- I have seen every video about Hedera Hashgraph, the official and unofficial ones. As far as I know, it's only viewable in the source code which is available on Github. A developer should be able to make sense of it. You could also run the SDK and see how it runs in real time. As a separate resource, they also have a Discord for developers.

[@danieltong7900]

I have one question, what if during one small round, some user just drop the network, will that cause any issue? From my understanding, this algorithm works base on each nodes know how many computers are in one system, so how they communicate before each round and what if someone drop, will he/she just been considered as a malicious node and then out ?

[@Supremax67]

What you are viewing here is the permission version of the network. Permissionless has not been implemented yet, but when it is, they'll share which technique they use for that too. To be clear, Hedera's public DLT already has a plan for nodes connecting and disconnecting, we are just too early in the game to see non-permission nodes being used. The network is already doing 3 million transactions a day with no signs of slowing down, the volume would have to be way bigger before they need to expand their network. Stay tuned to the space. Change is moving fast.

[@EdwinLiavaa]

Bravo Dr. Leeman Baird

[@DoYouwantToMakeMoneyOnline]

How do we all use the hash graph? where do I go to sign up? and when I am on the hash graph will be able to see all events and the details of these events, for example, who bought a house, what price they paid etc?

[@Supremax67]

Many questions, no easy answers. It all depends what you are trying to accomplish as a developer. Do you want the transaction to be visible or encrypted? Is privacy important for you or do you require an audit log for regulatory purposes. If it is a developer question because you have an idea for a Dapp, I recommend their Discord channel. If you wish to reach out to other community members for a different conversation, I recommend their Telegram channel.

[@shumymikaball]

How the hash authenticates the node? It seems that it could be easily forged.

[@mouduge]

Quick question: what prevents an attacker from creating millions of new nodes and therefore controlling the consensus?

[@desireisfundamental]

the network will show which nodes were created, when they were created, where they were created and when they knew the information.

[@michaeld2345]

exactly, and currently it's flying under the radar of hackers. If it grew more and became 'profitable' the next wave of malware would target personal pc's to disrupt the network by spoofing the events. It will never work, but on smaller applications, Hashgraph might be good for offering a service.

[@ilsunnylo3562]

Digital wallets still get hacked. Proven it will not work for the time being.

[@larryg3446]

Good one. Hashgraph does not have potential to replace blockchain. There are many problems Hashgraph will run into. Few of them are here, Hashgraph nodes would never be fully updated, so basically no one holds the full truth. No one ever knows exactly what everyone else knows. Are you sure it can prevent double spending? what if i have 1 coin, then i spend it on one node and another node that hasn't yet synced with first, it wouldn't yet know that i have already spent it. You cant assume that all nodes are good and syncing happens instantly. Second, for example, Suppose there are 10000 nodes in total and in intentions to attack this network, I create 2 nodes that only talk to each other like crazy but do not communicate with others, so no one talks to them. It creates 10 gb of data on those 2 nodes. After it is done, all 10 000 nodes will have to update with this new 10 GB of data and network is DDOSed. Tell me how this is going to replace bitcoin blockchain where you actually have to spend few millions to attack it. I already know how to destroy this technology, which is bad.

[@user-rm3iu7eo9t]

Anonymous love this

[@neurophilosophers994]

Is it possible to introduce a new computer or set of computers at the beginning of a round and can an address be a computer instead, meaning can one computer have many addresses associated to it which are each a vertical line in a hash graph?

[@neurophilosophers994]

I’m assuming for the latter yes because it is in a demo and for the former the answer is yes as well because otherwise I don’t think MMO gaming would make sense here. So I’m assuming yes unless someone can tell me different?

[@Supremax67]

@Salim Huerta -- They do have a Discord channel for developers if those are the type of answers you seek, as for the rest ... This video was part of their private ledger, all computers are on a permission network, I believe you might be inquiring about their pubic network which was announced in March. In a public setting, you need to stake coins to have a node, no coins, no nodes. The value of your node is determined by how much coin you have staked, however, you can't gain control of an hashgraph by staking more coins since they will simply move you to a different shard with nodes of similar weight. You can use MMO if you are willing to forego the consensus state and just grab the current state. Since most MMOs uses permission networks, it is not difficult to have people uses authentication before logging in the platform. Furthermore, failures of consensus reached could generate a report with the proper DApp running on top, if a single node is often flagged has not contributing to consensus, that could be an early sign of someone trying to hack the system. So instead of waiting days, weeks or months before a hacker is reported, the hashgraph consensus could flag them in a matter of minutes or seconds.

[@sheepsus]

thank you Alice and Bob

[@zzzamboni26]

If I understand correctly the main purpose of the hashgraph is to reach consensus on events(transactions) order. But what if Alice intentionally keep changing time on her computer (affecting events/transactions timestamps) while still remaining famous in next rounds and affecting average timestamp calculations for events/transactions ordering. I've read the white paper and haven't found answer to this question.

[@martinzitter4551]

Bob, Carol, Ted, & Alice -- SwingerGraph.

[@themoneyslaps]

Thank you💎

[@user-ff6on4hd3k]

it might be difficult to find flaws until there's a really big network. How many connections would a node maintain for both sending and receiving events in a million node network?

[@VenturiLife]

A wonderful invention.

[@kctheservant]

Thanks for your effort, and after the first hour I spent on this video, something on my mind: Comment: you have put rounds like based on time on slide 8, seems those events in that time period belong to round N. But turns out the result is not the same on slide 28. This makes things a bit confused. Question: (1) Why A3, B3, C3 and D3 know they are witnesses on round 3? Who puts the round boundary on slide 8 such that these events know they are the witnesses on round 3? (2) Why A3, B3, C3 and D3 would check A2, B2, C2 and D2, but not other events? Who tells A3, B3, C3 and D3 to check A2, B2, C2 and D2? Thanks again.

[@Supremax67]

You might want to try their telegram for your question. If you are a developer, they also have a discord channel.

[@mittlambert7765]

This is super cool. Will there be a war between blockchain and hashgraph? How can I invest in this?

[@georgeschannel3364]

Can anyone tell me if this company is going to have an ICO? If so, when does it look possible?

[@aminesila9276]

Great video

[@Supremax67]

Simple explanation guys, not shortest explanation. People need to understand the difference.

[@enasbate2625]

Hi, I was looking for your comments to contact you, I liked the way your replied and the belief you have in Hedera, I don’t know if I could reach you by email or ig or anything to contact you I would love to have a little talk about Hedera if thats doesn’t bother you

[@Supremax67]

@@enasbate2625 -- I am on their discord with the same username. Hedera Discord link is found at the bottom of their webpage.

[@JeyPeyy]

43:47 "C2 can't see the black event" Are you sure? It looks like it can see it through the gray event above A2 and A2 itself. Am i wrong?

[@PesarTarofi]

where do we buy it

[@moses5407]

How are rounds created? Timestamp top the explanation, please. I get how events are assigned to earlier rounds by events in later rounds but how are the rounds themselves defined/delimited?

[@paulmadsen]

its more that events are assigned to rounds, and the criteria for that assignment depends on the structure and denisty of the hashgraph. For an event to be assigned to a higher round (than its parents) the test is whether the event in question can 'strongly see' more than 2n/3 of events in the round of its parents. The round of the genesis event is defined to be 1, as later events are added to the hashgraph they are first assigned to 1, then 2, then 3 etc etc.

[@DrH5N1]

Bumping Hard Godspeed

[@tbuyus8328]

i had to read the comments to discover the white paper was a pre-requisite to understanding this. Maybe mention this in description. So I'm thinking this was for laymen and waiting for meaning to coalesce, but having too many layman delayed questions by 40mins to not feel I'm deluding myself that I truly understand. Then I stop and find pre-requisite and realise I do need to know the terminology and not just guess its meaning from normal usage. [So can I ask a meaningful question, probs not. But here goes: how do events know how many total previous events (witnesses?) there are in order to decide what the majority threshold is, or is this a majority of events visible to themselves and not the actual number of events? ]

[@maaifoediedelarey4335]

Yes, and perhaps take out the word 'simple' ! It seems to me that the 'consensus' they pride themselves on, is only re the order of events, not consensus about the accuracy of all other facts of transaction, as is guaranteed by longest block confirmed in bitcoin blockchain. If that's the case, then the Byzantine has not been solved.

[@MatthewMarcus]

Thanks for the explanation. Overall the system makes pretty good sense. A few questions and clarifications: 1) I noticed that every event in your example (except each member’s starting event) has two co-parents (itself and another event). I found this strange since a member could surely create events along the way which would only have a single parent (itself). Is this correct? 2) I don’t fully understand how rounds are created and determined. Towards the end of the video you spoke about events being received by famous witnesses, but you didn’t tie it all together. Additionally, you showed two grey events which would not be received by round 2 and likely would be received in round 3, however earlier in the video you showed them to be in round 1. Very confusing. 3) At minute 43:47, you claim C2 can’t see the black event. It can indeed see it via Alice, but I get since it’s not a famous witness it doesn’t matter anyway.

[@RishabDhar1988]

This is flawed system. Please see my comment above. With no incentive to mining, no one will mine the coin, and the number of nodes will become highly centralized, if only governments run them (happened in Venezuela). And now since they also dump the history, you have no way of knowing whether all those centralized nodes colluded

[@weneedtogoback8553]

-_- It's not a cryptocurrency...

[@S1UMLORDpaulo]

every event has recourse in the hash records

[@RishabDhar1988]

But if the nodes collude together, they will automatically agree to a common hash - irrespective of what it is. So no, without mining and history of transaction, this system is flawed. The whole system can be undermined by centralized colluding nodes.

[@RishabDhar1988]

There are no hash records to verify (if SHA-1 is broken - which it already is). So the whole argument of not finding a collision is a fallacy. The recommended SHA hash now is SHA-3.

[@adam-a-pa]

Geez man. I really got more from the thumbnail than I got from the entire hour long video. P.s. Bob should gossip less.

[@richardshandross3453]

Very interesting indeed. Thank you very much for the extremely clear explanation, including the tech reports (which I haven't yet read). One question: does this protocol deal with the double spend problem and, if so, how? With everything taking place so rapidly, and in such a disjointed manner - event X doesn't know which of its roughly contemporaneous other events will be famous, and is not checked against them - it's not clear to me how to avoid the double-spend until after it's already out the door. The graph doesn't reject events based on this factor, so there needs to be a ledger maintained somewhere ... sounds like every node will be constructing their own on the fly. Lots of points of failure unless everyone, perhaps, uses the same (or highly vetted) ledger-forming tools.

[@larryg3446]

It cant really prevent double spending. there will be no node that contains 100% of data. Syncing between nodes does not (can't) happen instantly.

[@TakeyMcTaker]

How is this not Paxos over gossip looking at the log backwards? Why not break ties or excess rounds with Raft?

[@Supremax67]

Its not and it wouldn't be, Hashgraph is patented which means if it was the same thing, Paxos would be in violation of a patent. I cannot pretend to know about Paxos so if you got some easy video tech explanation about it, I wouldn't mind comparing the two and seeing what are the differences.

[@brunospasta]

Wouldnt very qickly the amount of data sent for each event be very big? A lot of redundent information is sent through the wires.

[@drivn4895]

I love this

[@FlowScholars]

I'm curious about the scalability of something like this. How much of the information history has to be stored by each node over time? It seems like the speed is incredibly good here but if the storage explodes combinatorically with the number of nodes, that would make it unviable for any sort of global decentralized networks. I'd also like to see better explanations for the either how you deter or detect cheating. Great explanation of the mechanics though. Will there be any future explanations of the byzantine fault tolerance, and any sort of incentive/disincentive structure for cheating.

[@asabovesobelow8901]

Keagan McClelland this is what I want to know. What r the memory reqs? Guess I gotta read his papers

[@tidragos]

I would like to know this too.

[@user-bc1td4wo1h]

Yes, I need an answer too. In this video we see only few nodes. But if we have thousands nodes, how I need to send and store ALL they events? Calculating for all events will be slower to.

[@juri8723]

@@user-bc1td4wo1h one hint: mirror nodes

[@ThisOneCanSee]

Watched the whole video... So simple! It's like, why didn't I think of that?

[@justincase8787]

Simplicity is the ultimate sophistication ... Leonardo Da Vinci ;)

[@TxFw]

What if a supermajority isn't reached in a round? Does that mean the votes can't be counted?

[@TxFw]

So theoretically could a round go on forever or is there a limit? I'm also trying to understand how long syncs could take once you start to hit massive scale.

[@Blatt.online]

It's the Iron Bank of Bravos. You can remove one stone, another will just come in and replace it. Transparent, Flexible and Unstoppable

[@AlexandruBudisteanu]

From your lecture, I have understood that Hashgraph is not a blockchain Prof-of-Work solution, but rather an Acyclic Graph data structure with a Proof-of-Consent/Gossips protocol. Moreover, I have understood from your explanation that Hashgraph creates one or even multiple Acyclic Graphs in the network like IOTA (you mentioned that some of these gossips don't need to be sent to everybody else or to new nodes who will download the previous gossips; moreover the gossips may also arrive different points of time in the network) and because of this, the algorithm can't reward miners (full nodes) or even take fees. So, can Hashgraph reward miners and the miners can take small fees? Have I understood the algorithm wrongly?

[@AlexandruBudisteanu]

So, no Mining Rewards, but, my question is how can Hashgraph protocol be used in any ways to distribute transaction fees (gas) among the nodes(workers)? I personally don't see any ways you can distribute the transactions fees through out the network. Thanks.

[@AlexandruBudisteanu]

I believe that these "forgotten" facts that it is impossible to facilitate a Mining Rewarding system and Distribution of Transactions Fees in Hashgraph should be mentioned in every "lecture" on KZbin.... A mention in a panel session is something different that a formal and proven solution for enabling a concrete Distribution system for Transaction Fees using Hashgraph. I truly doubt that the current Hashgraph solution can facilitate a Distribution System for Transaction Fees.

[@benyaminewanganyahu]

Alexandru, could you please explain why you doubt that Hashgraph can facilitate a Distributions System for Transaction Fees? Couldn't the protocol include a small reward for sending each gossip that the person receiving the gossip can check? Please enlighten me.

[@LaszloPappBerlin]

Look at the crypto economics model of Hedera recently presented from Leemon Baird at the Hedera 18 conference in Dallas. I think that should also be online soon. That explains everything. In Hashgraph, there is no mining and with that there is no necessity to eat up Irelands energy needs just to artificially slow down the network - which I see as a clear advantage compared to PoW systems. Still, Hedera has its own crypto economics model carefully balanced to create monetary incentives to all participants in the network to actively participate in running the Hedera platform: be it a client, a node, a mirror node, etc.

[@DerekDominoes]

How is this different from IOTA? Seems very similar.

[@HederaHashgraph]

It is actually completely different than IOTA. The only similarity they have is that they are both built on DAGs, so it visually looks similar. Leemon goes more in depth in the differences in this video: kzbin.info/www/bejne/l4C8gWNrfNx-ec0 Thanks for the comment! Join us on telegram, t.me/hashgraph

[@robfielding8566]

this is a well known CRDT. everybody that needs to deal with scenarios that raft/paxos can't handle ends up here. example: one incrementing counter via max. then hashtable of counters per full node with increasing transaction ids. it allows for temporary partial orders that get resolved during gossip.

[@eddieschneider1947]

will there be a hashgraph coin on crypto exchanges?

[@HederaHashgraph]

Hi Eddie Schneider ! Thanks for the comment. Currently, we are only deployed in private, permissioned networks, so no public ledger, cryptocurrency, or ICO. If you want to stay up to date, check us out on Telegram: t.me/hashgraph Newsletter: hashgraph.com Forum: hashgraphboard.com

[@WoodwiseJoe]

I can see how this could be beneficial when compared to the possibility of corruption from an individual blockchain miner selecting transactions to include in the block. I would have preferred to see this software listed as open source. Plus, *"Trust me, it works."* is right up there with, *"I'm from the government, I'm here to help."* on my list of least trustworthy things people can say.

[@pr4nk5tr]

How does it prevent spam attacks?

[@Supremax67]

I believe you are referring to Denial of Service attacks. They already have plenty source material out there, let me know if you are still looking.

[@vinnyguardino3923]

16 more days!

[@kariuki6644]

why does it have a ceo? other bitcoin doesn't have an owner

[@manuelrothschild4705]

Doesnt decentralization imply no mean of advance as far as the technology is concern. So if there is no central party earning, why would they advance the technology...

[@greensyco]

If this works then great, I have seen numbers of 50k TPS and 250k TPS, but even then 250K TPS is not enough, PoS algorithms are talking about 500k - 1million TPS in the future, IOTAs tangle talks of much higher. In anycase, even if this works and is scalable, I see there is a patent on the technology, which is contrary to the ethos of decentralisation and fair distribution of value inside the community that adds value to the network. It may survive in the permissioned environment for a while (like the Intranet did) but the future is a public network and opensource... no patents.

[@TxFw]

Hashgraph what post-sharding TPS can we expect? And what do you mean by the importance of access pattern?

[@jacksableng4775]

I think hashgraph has exponentially more potential than blockchain. If only this open sourced or at least licensed this will surely gain more adoption. I hope the development team don't keep this technology to private application.

[@jaybrrr994]

exactly right

[@jacksableng4775]

You're talking about 1 million Tps on PoS, that might already include Multi-layer off-chain/side chain solutions, sharding and compression like lighting and plasma. Well the problem with all those solutions are that they are just a deferred consensus, add risks and add unnecessary complexity. All those throughput achieved by compression and are pushed into the main chain in a lump sum after many-many blocks confirmations. That's means the transactions might not be final until after it get a main chain consensus. In Hashgraph all transactions get consensus in milliseconds. So that is a clear advantages over Blockchain. And that doesn't takw into account ddos over main chain? While a Quantum hack is hacking the main chain which might even took days to reach consensus under heavy load. Blockchain is very vulnerable to ddos and the longer it takes to get consensus the more time it opens to Quantum hack.

[@enki-not-enlil]

Well if its not open sourced it won't be for use anyway. A decentralisation technology without open source doesnt work.

[@nguyenToZZ]

🎉🎉🎉

[@officialesteban4603]

The fact that the price of a hbar coin 10X since this video was published

[@U2Berror]

wow, it's like saying how quickly can you count and you're like by the time I reach 4 it doesn't even matter because now we know.

[@aaronhedrick2648]

HH: Hello future! Saft 3 investors: GFY!

[@commercialbrokerssicuro6667]

Im not developer, it seems very hard to understand for me ...sad!

[@IanMedeiros]

What prevents a malicious user to replicate bilions of nodes and broadcast gossips to flood the network infrastructure?

[@HederaHashgraph]

A sybil attack. We've written an entire whitepaper on this! www.swirlds.com/downloads/Swirlds-and-Sybil-Attacks.pdf

[@bejee45]

Stake -- Money

[@stackachusetts3233]

What about daylight savings?

[@Supremax67]

This doesn't apply to UTC which the nodes could use instead of their local time zone.

[@rishankkumarsingh]

Hbar is not pumping 😭

[@matttemar]

The future is here~!

[@tonybits369]

[@Cinnabuns2009]

This is like block chain + what we're learning from quantum physics... i.e. the future.

[@HederaHashgraph]

Thanks Cinnabuns! Join us on telegram! Http://t.me/hashgraph

[@HederaHashgraph]

Thee trolls are strong in this land - careful you must be. Newsletter on hashgraph.com is grand - and it charges no fee.

[@3harath]

this is basically the similar concept of Raiden network which works on eth blockchain

[@swbratcher]

@Hashgraph Could this distribute a huge pile of information that evolves over time and needs to be maintained in a healthy and accurate way for the benefit of all who want to access it? Think Wikipedia. What a noble first endeavor that would be for a technology like this.

[@HederaHashgraph]

Hi swbratcher ! Thanks for the comment. Yes. Hashgraph serves as a consensus algorithm, so you could certainly build a file system on top of it. If you want to stay up to date, check us out on Telegram: t.me/hashgraph Newsletter: hashgraph.com Forum: hashgraphboard.com

[@nickfernald2092]

Okay. Basically un-hackabe, But how do we know it will grow? This is very confusing to the normal person. How will you explain this to someone that barely believes in bitcoin? How will people using hashgraph for their businesses fluctuate the circulating supply of their coin?

[@rvnmedic1968]

Before even watching this and others purporting to explain the Hashgraph, one needs to know and understand all the euphemistic terminology. What is consensus, what are nodes, what are members? So, I now need to get good definitions of all the terms before trying to understand this and/or the Blockchain. No wonder the average person on the street knows nothing about cryptos and doesn't want to know when they encounter the crypto world.

[@Supremax67]

Unfortunately the crypto world is filled with people shilling and spreading FUDs, so it is hard to get good information from emotional investors. I been following this project very closely and it does show a lot of potential. You could try their telegram channel.

[@MrLawman82]

What's holding this back, why isn't this mainstream or is it?

[@Supremax67]

Depends your definition of mainstream. If you are referring to mass adoption, well, that only occurs when it is too late to invest. If you are referring to brand recognition, that has already happened and is growing each day.

release the code & get it battle tested, what I question mostly that no more than 1/3 of nodes would be vicious, mostly those famous one...

[@HederaHashgraph]

Hi Robert! Thanks for the comment. We are testing Hashgraph now through our first PoC - CULedger, a consortium of over 6,000 Credit Unions, who collectively serve over 105 million members (Think R3/Corda for the big banks). You can check out their project at culedger.com Full results will be coming in the future. If you'd like to follow us, Telegram: t.me/hashgraph Newsletter: hashgraph.com Forum: hashgraphboard.com

[@pip0109]

DO THEY CHEAT! ARE YOU TESTING CHEATING?

[@ZenfulLifeTV]

I was reading your privacy policy on your website. This is in no way a private service.. it's a shame. It mentions a lot about collecting personal data and third parties involvement... it makes me very weary :/ ..

[@Supremax67]

Are you referring to KYC? Because KYC is a standard regulation to deter from fraudulent activities; its purpose is to prevent money laundering, terrorist financing and theft. Important not to be confused with the privacy of the network vs the anonymity of an individual. You can't really hold their company accountable for following the law & regulations. That makes no sense.

[@ZenfulLifeTV]

It wasn’t the KYC. There is a section in the privacy policy: “ For example, we collect information about your device and its software, such as your IP address, browser type, Internet service provider, platform type, device type, operating system, date and time stamp (a unique ID that allows us to uniquely identify your browser, mobile device, or your account), and other such information.”

[@Supremax67]

@@ZenfulLifeTV -- That privacy policy is no different then when you are surfing the web and ask you if you accept cookies. That is typical website questions. You make it sound like this is something new, but in fact, about 90% of the internet already does this, some does this and does not notify you about it. They are many reasons to do this, first instance, browser/device supports. It is expensive to support multiple old browsers/devices over time, like any sites, they need to know if there is a need to keep supporting those same devices. IP addresses being tracked is nothing new and in fact encouraged, not everyone is a good actor, some people may decide to flood their site with packets, making their site crash. They need to know which ip they need to block so that service can continue for others.

[@upgrade1583]

Here's my simple explanation. Take same tangle diagrams from IOTA and convert them to 2d. Tell everyone its to complicated to explain. Market it as something new.

[@epm8593]

When PhD says "simple".

[@trenarhodes6631]

I talked to some key eos peeps about this seams they are not impressed I'm sure that ADA Cardano IOHK would say the same, might scare Euthereum though, but still I will keep an eye on it...

[@HederaHashgraph]

Hi Trena Rhodes ! Thanks for the comment. It makes sense for them to feel this way as they are all public ledgers and we are only serving private, permissioned networks. The security concerns are much different. Here are 3 great sources to stay up to date! Telegram: t.me/hashgraph Newsletter: hashgraph.com Forum: hashgraphboard.com

[@slomnim]

could you describe a real-world / layman application of this technology, even if it was just on a private network / backend? what are the most applicable real world uses for this specific technology, and can it ever -become- a public "cryptocurrency" of sorts? Is there an application you're working on now with this development? as well, how private does this type of distributed computation end up being? or say, could one make it cryptographically secure using this coding and yet also invisible transactionally if it were to be used as a currency? this seems to have potential for creating an actually fungible currency if it was kept private / secure effectively.

[@ArtMaknev]

Raiblocks looks very similar to hashgraph.

[@hashtech7487]

Hashgraph just a form dag technology like iota, nano, and bytball! They all just uses a different algorithm

[@hashtech7487]

They all use gossip to gossip protocols!

[@Supremax67]

But Hashgraph is not a blockchain, that is a poor comparison, it's fundamental function of Hashgraph is also different than what IOTA, Nano and Byteball is trying to accomplish. For now, nothing on the market works like Hashgraph.

[@harleyquinn8202]

It is not Sybil or spam resistant

[@HederaHashgraph]

Hi Harley! In permissioned use cases, you do not need to worry about Sybil attacks. However, if you want to theorize about a public ledger (we do not have one), you may want to check this out: www.swirlds.com/downloads/Swirlds-and-Sybil-Attacks.pdf