When a company says that its technology is going to revolutionize not just blockchain but also the internet, it’s going to garner a lot of outside attention. It’s also going to attract a lot of scrutiny – much of it likely negative (given the sheer number of entities that have spent the last twelve months or more building platforms on blockchain technology.
This claim is one that a relatively unknown company called Swirld’s is making.
The technology that it feels can live up to this claim is called Hashgraph and it’s essentially an entire reimaging of the processes and purposes that underpin bitcoin and blockchain technology in general.
Unfortunately, much of the coverage that Hashgraph is attracting right now is superficial, in the sense that it highlights the features that set Hashgraph apart from, say, the blockchain that underpins bitcoin, but doesn’t fully get to the bottom why these features matter and – in turn – why the company behind Hashgraph is confident enough to make the above stated claims.
With this in mind, we took a deep dive into the Hashgraph concept in an attempt to answer the question – is this really something that has the potential to unseat blockchain?
Here’s what we came up with.
An introduction to inefficiency in blockchain technology
Before we get on to Hashgraph, it’s probably worth spending some time explaining why something like bitcoin’s blockchain might need improving on.
Basically, it’s incredibly inefficient.
Not inefficient in the sense that it doesn’t do its job properly – bitcoin’s blockchain does an excellent job of recording bitcoin transactions in a time stamped and reliable manner – but instead in the sense that to achieve this outcome, a huge amount of energy is expended. And we’re talking literal energy here. Huge amounts of electricity are required to power the mining technology that solves the problems that are required to mine a block. That’s a problem for a couple of reasons. First because it’s going to lead to the migration of mining technology to the geographical regions in which electricity is cheapest – right now, somewhere like China. If China let’s all the miners set up within its borders then overnight decides to nationalize the miners in question, it could very easily bring bitcoin down.
The second problem (and likely the more important one right now) is scalability. Bitcoin transactions are limited to around 7 a second right now. That’s nowhere near enough to meet the demands of the global credit card transaction count (somewhere around 7,000 a second) alone.
Scalability can happen, but it’s tough to achieve without community consensus – as the recent Segwit2X cancellation has illustrated.
Further, and perhaps counterintuitively, bitcoin’s blockchain is designed to slow itself down. The more miners that come on board, the more difficult mining becomes, so as to limit the time it takes to mine a block to one block every circa 10 minutes. This limit is in place to virtually negate the chances of two blocks being mined at the same time and to – in turn – reduce the chance that one block’s worth of energy is wasted (bitcoin’s blockchain will, in the event of two blocks mined at the exact same time, favor one and nullify the other).
So that’s what’s wrong with a public ledger blockchain like bitcoin.
Now let’s look Hashgraph
How does a Hashgraph (we say ‘a’ Hashgraph here as the term is comparable to a blockchain, in the sense that it’s the name of the dDLT) aim to solve these problems?
Right now, public blockchain systems like that of bitcoin rely on consensus proof of work to decide which order transactions took place. The more confirmations received, the more certain the network can be that the order is correct and the higher the degree of consensus.
In contrast, on a private blockchain, a leader system is used. Basically, all of the participants in a network send their transactions to a leader and the leader decides the node. This is a neat solution to the inefficiency of a public blockchain (the inefficiency that’s outlined above) but it opens the blockchain up to security issues – primarily, a DDOS attack on the leader.
With Hashgraph, the system used to arrive at consensus is called a gossip protocol. This is a pretty well known computing term and it basically means that every time a transaction takes place, it’s shared with all of the participants in the network very quickly. Say one computer learns about a transaction. That computer sends it to another computer, which then sends it to another. At the same time the second computer is sending the transaction to a third computer, the first computer is sending the transaction to another random computer on the network. This allows the transaction data to spread incredibly fast (exponentially) and very quickly everyone knows about the transaction.
But there’s an extra element and it’s in this extra element that the Hashgraph advantage really relies.
Gossip protocol isn’t really anything new. With a Hashgraph, however, the sending computer adds two bits of information to the piece of gossip (the transaction) it’s relaying to another computer on the network – the name of last message the computer sent and the name of the last person who sent a message to the sending computer.
This extra bit of data (well, two bits) is incredibly lightweight but can also basically allow for complete transparency, in an instant, and by proxy complete consensus and confidence in the voting system that generally underlies these sorts of networks, without having to expend huge amounts of energy mining blocks, confirming transactions, all that sort of thing.
With a system like this, not only can transactions be confirmed far faster than with a blockchain system like that of bitcoin, they can also happen at a much higher rate – circa 300,000 transactions per second.
At the same time, the system is far more secure than any of the current blockchain iterations.
There’s a security characteristic called Byzantine fault tolerance, which basically means that there’s a time at which the network can be 100% certain of the order in which transactions on the network took place.
Bitcoin and blockchain technology is often said to be Byzantine fault tolerant, but it’s not. The higher the consensus, the closer the network is to certainty, sure. It will never reach 100% certainty, however, and as such, cannot be called Byzantine fault tolerant.
Hashgraph is Byzantine fault tolerant, with this tolerance achievable using the two bits of extra data attached to each sent-out transaction.
So let’s get back to our initial question – is this really something that has the potential to unseat blockchain?
From an adoption perspective, it’s impossible to say. As mentioned, there is a large amount of time and money invested in blockchain systems right now and this money is going to be resistant to change.
From a technical perspective, however, Hashgraph seems vastly superior to any of the distributed ledger systems currently in use.
And if there’s one thing of which we can be pretty much certain, technical superiority often has a way of overcoming market and participation-fed resistance.
David Cohen – a key endorsement
Further, it’s not just the creators of Hashgraph that are standing behind the claim. Many reading will likely already be familiar with David A. Cohen. For those that aren’t, he’s the known the world over for his work in the field of decentralized software and – more recently – blockchain technology and application. Prior to this, however, he was a big player in the artificial intelligence (AI) space, in particular as relates to autonomous task completion and resulting machine learning.
Cohen founded Infotility, an AI company, and created what he called GridAgents, which was then applied to smart grid infrastructure to form the basis of an application designed to create smart electrical grids in major cities (one of which he modelled was NYC).
Anyway, prior to blockchain, this sort of self learning and autonomous machine learning was missing a vital component – a gap that blockchain filled on creation.
Cohen picked up some funding from the US Department of Energy and set out on a mission to try and figure out which version of blockchain technology would fit best with his existing AI model and – over a discussion one day with a former acquaintance – was introduced to Swirld’s and Hashgraph. After a deep dive into the technology that underpinned this version of blockchain, as part of which he uncovered the features and differentiators we’ve outlined above, not only did he accept that Hashgraph was ideal for what he was trying to achieve (that is, to create the next generation of industry, what he called Industry 4.0, where autonomous learning and AI technology dominate the framework that underpins operational activity), he joined the team himself.
Image courtesy of Mike Seyfang via Flickr
Disclaimer: This article should not be taken as, and is not intended to provide, investment advice. Please conduct your own thorough research before investing in any cryptocurrency.