Hedera Hashgraph – NYC Meetup – May 14, 2018 (Kyle Armour)

– I'm Director of Community
for Hedera Hashgraph. I'm gonna walk you through
kind of what Swirlds is, what the Hashgraph algorithm is, what Hedera is, and what
we're doing in this space, and then the Carbon Money
guys will be here shortly. They are developing the
stablecoin on top of the Hedera public ledger. So, first off, I'd really
like to talk about this. I'm pretty active in our Telegram channel, and a lot of people think
that, just in the last week, that we have the same
vision as Satoshi Nakamoto, which is just really not true. Leemon's vision came from
a very different place than what he was going for, but he says this in a series
that we have on YouTube. "We believe in a future
where you can wave your hand "and create a shared world for free. "It's easy, trustworthy, reliable, "and you can invite whomever you want "and share whatever you want. "You can collaborate and create
movies, music, documents, "3D objects, and converse
without dictatorship "and in complete fairness.

"A system that we can trust
going to work the way we want, "and if my computer dies
and I lose all my data, "I can know the data is still safe. "No servers, complete trust
and security, and fairness." So, you can kinda see, in this model, I don't believe he talked
about cryptocurrency anywhere. Of course, we need a
cryptocurrency and a public ledger, but Leemon's vision was
much bigger than that.

So, what is Swirlds? Swirlds is a company that has patented and invented the Hashgraph algorithm. It's a developer of a next generation distributed consensus technology. We're based in Dallas, Texas. I think we're just, I mean,
a bunch of us are remote, but we're somewhere over 30 employees now. It's scaling pretty rapidly. We incorporated in quarter four, 2015. That's when the algorithm
was actually invented. The last two years, we've been, you know, testing this in permission
private networks, so it's not like your
traditional white paper going straight into a public ledger. This algorithm has been
tested in a permission setting for quite a while, and we
had three million raised on behalf of Swirlds led by
New Enterprise Associates.

We have since, March 13th in
New York City in Times Square, we announced the Hedera public ledger that we were intending to do, and we have raised 18 million for Hedera. On this right here, you
have kind of the generations that we believe everybody
kinda has their own, you know, generations that they go by, but we kinda see ourselves
in this fourth generation. Of course, cryptocurrencies
came first in 2009 with bitcoin. Shortly after that, there was
private permission networks that were doing ledgers. We can store cryptocurrency. Why can't we store
something like log it trails or property deeds or things like that? Then, of course, in 2015,
we had Vitalik Buterin and smart contracts, so
if we can store those two different kinds of assets, why
can't we do a simple contract to facilitate the actual swap? And then, the last thing
here is fair markets, so what do you need to do fair markets? You need time stamping,
ordering, and all of that.

A little bit about the founders, you have Mance Harmon and Leemon Baird. These guys met in 1993, working for the senior
scientists in machine learning. They both taught at the United
States Air Force Academy, both in cybersecurity, Mance kind of more on the leadership side. He has a Masters in
Computer Science from UMass. Leemon, you know, is an
extremely brilliant guy that holds a record at Carnegie Mellon for the fastest PhD in
Computer Science there, two years and nine months, and he has over 100 publications
and peer reviewed journals on computer science, machine
learning, and mathematics. Together, they have formed two companies.

This is their third one, and they have sold both
of those companies. So, we'll get a little bit
into the actual technology, what is Hashgraph? So, Hashgraph is not a blockchain at all. It's actually a DAG, a
directed acyclic graph. You can see what that looks like. Here's your linear blockchain, here's DAG. It's a very simple, elegant algorithm.

It's mathematically proven, and it's implemented in software already, as I kind of mentioned before. It already exists. You can go online and
download the SDK right now. And it's also the only bank-grade
consensus in the market. What that means is we have
a property that's called asynchronous byzantine fault tolerance, and it's the strongest form of consensus you can actually have. So, quickly, what I'm
gonna do to kinda just grab a level, how many
people actually know of Hashgraph and are pretty familiar? Okay, about half of you guys. That's pretty good. Okay, so, what I'm gonna
do is kind of show you the last eight years in where
this technology has gone. We're gonna look at the
differences in performance and security per punch, so again, this is kind of a table of contents.

We're going to jump into
leader based systems, proof of work blockchain, economy based systems,
voting based systems, and then kinda show you how
Hashgraph comes in after that. So, leader based systems, these can do about 1000
transactions per second, so they're really good for POCs, but they have a very
fundamental security flaw in that these are all the
members of the networks, so if you can imagine this is a bank, all of these members have
to depend on the leader to organize these transactions
and send them back out to the members, so if one of these members is compromised by a virus of malware, or maybe just a disgruntled employee, and it can actually direct
a denial of service attack.

That's where you basically
spawn up a bunch of computers. Maybe you have compromised
baby monitors or printers, or you just go to a digital ocean and just buy a bunch of
servers, whatever it may be, you can then direct all those
computers at this leader by sending the IP address
out of the network, and you can do that in
full anonymity because it's just a couple bytes, you
know, it's just a few numbers. What will happen when that
takes place is that this leader will actually go down, it'll go offline, and by design, a new
leader will be elected, and you can't tell who's
actually compromised, but they have to know the IP address of that leader by design, so this computer, this malicious computer, just redirects that
attack, that DDoS attack.

So, it's very susceptible to DDoS attacks, and these are used a lot in
permission based networks where you don't have such secured flaws, but it's not the hundreds of thousands of transactions per
second we're looking for. Good for the POCs, as I mentioned, and it has kind of a big security flaw. So, proof of work
obviously came with bitcoin and Satoshi Nakamoto's vision, so what happens here is that a transaction is actually submitted into the network. These miners then fight over it, extremely difficult cryptographic puzzle, basically guessing numbers
until they get the right number.

When they guess that number,
they then announce it to the community and they get to add that block onto the chain. The problem here is that it's
extremely slow by design. Actually, the developers have
to make it more difficult, the puzzle has to be difficult enough that the system is slow enough that it doesn't spiral out of control. If it was going too fast, you would end up getting a hydra head, and too many people would
be putting on blocks. The other thing is that
it's extremely unfair, so the miner that
actually wins this puzzle gets to decide where
transactions go in a block. They get to decide what
transactions actually go inside a block, so they could be bribed to just ignore your transaction, so you can say it's not fair, and again, it's extremely slow, but it is very secure
in that these systems haven't been hacked yet, though.

What a lot of people are
trying to do in terms of to solve this transactional
throughput being so slow, because bitcoin, I think,
is somewhere between five, 10, 15 transactions per second, and Ethereum a little bit faster, a lot of people are moving to
these economy based systems. These are basically proof of something, proof of state, usually, and what happens here
is that there's kind of a fundamental assumption that is made that every node in this network will act in their own best interest, that they will do whatever they can to make the most money. It's a very bad assumption to make, because it doesn't take into consideration things like viruses, and
people to act irrationally. Something we often say
is could you prove that the U.S.

Stock market
will never crash again? Of course not. You can't write a math
proof that says that, but if we're gonna trust a public ledger with trillions of dollars worth of assets, we should really try to have a math proof that says that a single attacker cannot do significant
damage to these systems, and they cannot actually
answer that question. So, we'll see, you know? None of these have really been
fully deployed in the space, but they are coming. If you go 35 years back
in computer science, there was voting based algorithms. These were like the fundamental, the very beginning of the internet. I think it's the way they
wanted to build the internet, but they couldn't actually figure this out back 35 years ago.

In these networks, you
would have every computer, basically, the transactions come in, they have to vote yes or
no on this transaction. Every computer would have to
talk to every other computer. It's basically N-squared work, extremely inefficient,
because every computer has to talk to every other computer. However, you can see, if
that kind of communication was taking place, you have a
very strong form of consensus. These are where the asynchronous byzantine fault tolerance comes from. These had the gold standard of consensus, so the question was how can
we do a voting based system without actually doing the
votes, the inefficient part? And that's what Hashgraph does. Basically, how Hashgraph does that is every time you're gonna talk to someone, you're going to include your own hash and the hash of the last
person you've talked to. That continues to gossip out, so I'm gonna talk to you, and then you're gonna go
talk to somebody else, and when you're talking to somebody else, I'm going over here and telling them that, and we all just spread through gossip, and after a certain amount of seconds, usually about two to three
seconds, we've proven, everyone will have a picture that's built up of the Hashgraph.

They can then look at the
Hashgraph in their memory and input that data into one
of those old voting algorithms, into one of these, and by doing that, we can solve that whole inefficiency and actually get some amazing things, so we can do hundreds of thousands of transactions per second. We have the gold standard, asynchronous byzantine fault
tolerance for consensus, and we also have new
properties called fairness, so we have fairness of order, timestamps, and we have timestamps on every event, so what basically happens
is every event that happens from a node, we also include, I should show you this in a second, what's actually in an event, but we have the two
hashes that I mentioned and the timestamp, and
the timestamp itself, you look at all of the events and you take the median
timestamp, not the mean, so we'll take the median timestamp of all the timestamps that we get, and we'll get a community timestamp, when the community
believes that that event actually took place. So, here, you have a
little bit of a summary, kind of what I just went
through with, obviously, Hashgraph being unique in the space.

Quickly, I'm gonna kind of go over again what the gossip looks like, so here, you have Alice
that has a transaction. She's gonna send that transaction to Dave, and then they're both
gonna tell two people who tell two people, and then everybody has the transactions. So, that's one way, right? What happens if two people
do it at the same time? So, Bob and Carol both
have transactions here. They both send two out. They send, again, exponentially going out, and there you go, and
then what you'll notice, some of the order's not right, so that's where the timestamps come in that I just mentioned. Here's kind of what the
Hashgraph actually looks like, so these are nodes, obviously, here. Alice, Bob, Carol.

These are all the communications
that are happening, and that builds up, and
that data right there is what gets inputted into
those full voting algorithms. So, gossip is basically,
let me define that, too. Gossip is just calling up a random node and telling it what you know. This is not a novel idea. It's used in bitcoin theory and other areas in computer science.

Gossip about gossip was the
breakthrough that Leemon had, and that was basically
additionally talking about the action being had, so
we talk about the gossip, but then we're gonna also talk about who we gossiped with, and the last person. And then, the virtual voting is, again, we can just run that voting
algorithm in our head. We can look at the data. Everyone has the same data, and we all come to the same consensus. So, there you go. You have the two hashes. The hash of yourself, previously. The hash of the last
person you've talked to. Whatever transaction may
be in there, payloads, and the timestamp, and of course, there could be dishonest members in there and maybe someone's claiming
they did their transaction in the year 2050, or someone else put their
transaction in at year 1900.

That's, again, why we take the median. Any of those dishonest
nodes will get kicked out. So, what are the new use cases? So, the three main ones that
we are really excited about are fair markets, fair
auctions, and actually games. You can, of course, do micropayments, and this would be the first, I think, really good demonstration of micropayments really being successful. Yeah, and I'm really excited,
personally, about the games. I think, in the future, we'll have games like World of Warcraft
really deployed on here, which will change the
gaming industry a lot, when independent developers
no longer have to. I think, you know, Machine
Zone is a company that's deploying on us, and they have, I think, a $200 million server
infrastructure for their iOS games, all across the world. When you can relieve
that, and not need to add a server infrastructure for gaming, you know, it's gonna be
a very different place, so very excited to see that.

Here is the SDK. If you want to download
it for your developer, you can do that now, Swirlds.com/download. That comes with, this is the config file, which shows you kind of the IP addresses. The SDK, currently, is used mostly for
permission based networks, so if you wanted to develop
POC, you could download this. You can actually test the
Hashgraph algorithm, too. We actually tested it using Raspberry Pis, which are $30 computers, so this algorithm is that
small that they can process. You know, don't need a massive
server rate to run this. In this case, you would
define all your nodes here, and you could test accordingly.

Comes with some demo
apps, so here's a game. Here's the Hello Swirlds, so
you can actually kinda see it coming to consensus. I know that's really hard to see, but you can see it coming to consensus and reordering things as it's going, and then we'll talk
about the public ledger, so that's like probably
the most exciting thing. So, obviously, I mentioned earlier, March 13th in Times Square,
we announced Hedera, which is, you know, the last year, it's always been a vision
for Mance and Leemon to do a public ledger, but they wanted to test
the algorithm first and know that that was working, and then they kind of addressed the issues that we see in space, so they
spent the last six months to a year kind of looking
at the public ledger space and trying to figure out the problems in what people were talking about, why this is not becoming
a mainstream thing. So, a couple of those
things were performance, security, governance, and stability. So, we believe that you
need hundreds of thousands of transactions per seconds of latency.

We need to eliminate
threats and ensure fairness. We need mathematical proofs
that guarantee our security. We need governance, so we
have to have, you know, experts for technical, financial, legal, all the regulatory things
that are gonna come about when we deploy these worldwide networks, and finally, stability,
so one of the things that we've seen is that, when you create, the open source communities,
obviously, we love them, and they are catalysts for change, but when you combine them
with a cryptocurrency, the incentives to thwart
and to do your own thing and to break off is so
great that it creates chaos, and we aim to solve
that, so what do we do? We have the hundreds of thousands of transactions per second. We have the asynchronous
byzantine fault tolerance, which we have the math proof of the strongest security possible, and we came up with a governance model that was 39 global blue chip companies. They do 10 billion plus in revenue. They will be the best
in class enterprises.

We will have two per sector, no more, so we might have two banks, we might have two insurance companies, we might have two telcos and
two fast food consumer goods. We believe that these companies will not collude with one another because of their separation in the sector. And lastly, the stability, so, again, we are a open review. Many of you probably know that, we are open review, so
it's not open source. What that means is
because we have a patent on the technology, that we're not gonna come after you if you decide to steal the code, but we're going to use
that to guarantee to, especially enterprises and
developers who are worried about spending two million dollars to develop a DAG on one
of these public ledgers, they wanna know that tomorrow,
it's still the same chain. They get really worried
when things are forking and they don't know what's going on, and they're just not
gonna invest in something that tomorrow could fork, and they don't know what's going on.

It becomes chaos again. So, one thing with that,
we are open review, but we do not require
permission from anyone to develop an app or
use the public ledger. It's completely public. You don't have to talk to us. There's no license. The software you develop
is completely yours. You never have to talk to us. It's a normal public ledger, and the source code will be
public review with Version 1.0. I think we're looking at a
quarter four public launch at some point in time. So, again, we have the TCP/IP,
the internet, right here. Sitting right on top of that
is the Hashgraph algorithm and Hashgraph SDK. It will come with three services, a cryptocurrency with native
support for micropayment, a file storage system, and the Ethereum virtual machine will be sitting right on top of it.

We've see that, you know, there's I think over a
thousand apps on Solidity now, on Ethereum, so why not just
plug that right in there? Yep, and that's it for me..

You May Also Like