What Is Hashgraph?
Cryptocurrencies and blockchains are heralded by many as the most potentially impactful technological advancements since the internet. But what if the blockchain technology used to create all of these cryptocurrencies is just a stepping stone to something better?
That’s the claim made by Hashgraph, a distributed ledger technology with many of the same applications as blockchain. Hashgraph is secure, decentralized, cost-effective, and can process transactions faster—and more fairly—than any blockchain.
Considering that, maybe the articles declaring Hashgraph to be the future of decentralized technology have some merit. Then again, they make some lofty claims, of which it’s hard not to be skeptical. Not to mention the fact that blockchain has nearly a decade-long headstart in both development and public acceptance. Even if Hashgraph is as good as advertised, competing with blockchain is still a monumental challenge.
With those points in mind, let’s take a deeper look at Hashgraph technology and go from there.
How Does It Work?
The figure to the left is a hashgraph, the data structure alternative to blockchain.
For simplicity, there are only 4 columns in this graphical representation. Each column represents a full node in the network, so you can imagine that there would be hundreds or even thousands of columns on an actual hashgraph.
Each circle represents an event, which is analogous to a block in the blockchain. In other words, events store data about transactions. Additionally, every event has the hashes of two recent events below it and is digitally signed by its creator. This enables Hashgraph to achieve cryptographic security in much the same way as a blockchain.
Every line connecting one event to another represents one node randomly syncing with another. When a node syncs to another node, it shares all of the events it knows that the second node doesn’t know yet. This syncing process is, fittingly, called gossip, and it continues forever as the hashgraph grows upwards.
The colored circles in the figure are special events called witnesses. It would take a while to explain how witnesses are differentiated from ordinary events. So, rather than do that, let’s just discuss what they do. Witnesses are the key events that carry out Virtual Voting as part of Hashgraph’s consensus algorithm.
The result of this Virtual Voting is that Hashgraph can reach consensus on the validity of transactions without costly Proof of Work (PoW) computation, and it can process hundreds of thousands of transactions per second. Ultimately, this is one of the biggest reasons that Hashgraph is a potential competitor for blockchain.
If you’d like to understand how Hashgraph works in greater depth, you can learn more by watching videos on the Hashgraph website or reading their whitepaper.
Comparing Hashgraph and Blockchain
Before diving into a comparison of Hashgraph and blockchain, it’s important to note that Hashgraph is not an open-source project. That means that the claims made about its abilities have not yet been independently verified. However, they are being independently reviewed now, and the Hashgraph team has stated that they are confident in the outcome.
One particular claim that raises eyebrows is that Hashgraph can process hundreds of thousands of transactions per second. One of the factors limiting processing speed is the state (i.e. storage) of the blockchain or hashgraph. As more transactions are carried out, nodes must store more information, decreasing processing speed. It will be interesting to see whether Hashgraph has a plan to sidestep this problem.
With that being said, Hashgraph is not attempting to get ordinary people to invest through an ICO, and doesn’t otherwise give the impression of having any ill intent. So, for the sake of this comparison, we will assume that its claims are valid, as there is no reason to believe that they are not.
Let’s start by looking at one of the most contentious issues in the blockchain space today…
Transaction Processing Speed
Bitcoin’s blockchain typically processes less than 10 transactions per second. That’s why Bitcoin’s viability as a currency for mass adoption has been under question, and why Bitcoin Cash was created. Second-layer solutions like the Lightning Network may be able to improve scalability drastically, but their implementation is still a ways off.
Hashgraph, meanwhile, boasts that it can process hundreds of thousands of transactions per second, making it the significantly better option for micropayments and low-fee, fast transactions in general.
Fairness
Another factor on which Hashgraph differs from blockchain is ‘fairness’. With blockchain, miners have the ability to choose the order in which transactions appear in the block that they mine. That means that miners can, in theory, manipulate the order in which they process transactions to somehow benefit themselves or harm a party that they dislike. Hashgraph simply doesn’t have this possibility, as they use Consensus Time Stamping to achieve fairness in the order that transactions are processed.
Mining and Sybil Attacks
Here’s where Hashgraph makes their big trade-off to achieve the outcomes discussed above. Hashgraph does not use Proof of Work (PoW), making it far less costly to process transactions relative to mining on the blockchain. However, this comes at the cost of security. Let me explain.
In peer-to-peer networks, one of the attacks that you might face is what’s known as a Sybil attack. This is where an adversary controls multiple nodes on a network by creating false identities, opening the possibility for an individual to gain control of a large percentage of the network.
Bitcoin is extremely Sybil resistant. This is because all miners in the Bitcoin network are trying to mine the next block and earn the block reward, so they are incentivized to use their full computational power. As a result, it’s impossible for an attacker to create more blocks on the blockchain by generating false identities and running more nodes, as they are still limited by their computation power. This is represented on the middle pie chart in the figure above, in which each new identity created by a dishonest node simply takes some of the computation power from an existing dishonest node without adding to the pie.
What happens when you don’t have block rewards incentivizing node operators to use all of their computing power? Quite simply, each node uses the minimum computing power needed to keep the network running. It’s possible, then, for an adversary to create multiple identities, run multiple nodes, and control a larger percentage of the network than the honest nodes. This is represented on the pie chart to the right in the figure above, in which each new dishonest node increases the percentage of the total computing power that’s controlled by dishonest nodes.
Hashgraph has not yet achieved the same level of Sybil resistance as Bitcoin, Ethereum, or other blockchain-based cryptocurrencies. Ultimately, that is one of the most significant reasons that Hashgraph will struggle to gain the trust of cryptocurrency enthusiasts.
However, Hashgraph has addressed the concerns about Sybil attacks, which you can read about here. Being less Sybil resistant isn’t the end of the world by any means, and Hashgraph still has tons of very useful applications at a fraction of the cost of Bitcoin. It just doesn’t have quite the same degree of security in this particular respect.
The Hashgraph Team
Hashgraph was created by Leemon Baird. Baird earned his B.Sc. in Computer Science from the US Air Force Academy before moving on to earn his Ph.D. in Computer Science from Carnegie Mellon University. He is also the co-founder and CTO of Swirlds Inc, which builds software using the Hashgraph consensus algorithm.
Swirlds Inc’s other co-founder is Mance Harmon, who also has an impressive background in computer science and tech entrepreneurship.
For more information about Baird, Harmon, or Swirlds, check out the Swirlds website. You can also get involved with Hashgraph through their Meetup.
Competitors and Challenges
Currently, Hashgraph has not released any plans for an ICO. In fact, it’s not clear whether or not Hashgraph technology will ever be utilized by a decentralized cryptocurrency. For now, Hashgraph’s funding comes from private enterprise businesses who may utilize the technology for centralized applications.
However, there is some precedent for a non-blockchain-based cryptocurrency. That would be IOTA, whose ‘tangle’ closely resembles a Hashgraph.
One obvious improvement that Hashgraph has over IOTA is that nodes can bundle transactions together, whereas each ‘event’ in IOTA is a single transaction. Therefore, Hashgraph’s throughput can be faster, while the bandwidth and storage space necessary for a new node to join the network is smaller.
For now, IOTA has no reason to worry about losing a piece of their market to Hashgraph. We’ll be sure to update you in the event that Hashgraph releases news about a potential ICO in the future.
Last Thoughts
Hashgraph does indeed offer some improvements over blockchain technology.
However, when it comes to computers, you don’t get big points for 20% improvements. You need an order of magnitude improvement to really shake things up. Given that Hashgraph is more susceptible to Sybil attacks than blockchain-based cryptocurrencies, it may not be the answer for fast, secure, and decentralized micropayments.
Furthermore, part of the value of blockchain is that it is relatively easy to understand, which makes it a more likely candidate for mass adoption and extensibility. Hashgraph would struggle more in this respect, as the technology is quite complex and takes a lot of time to understand.
That being said, the centralized applications for Hashgraph range far and wide. For existing private institutions, it may make more sense to implement Hashgraph than a blockchain, given its superior efficiency.
Whether Hashgraph ever becomes relevant in the cryptocurrency space remains to be seen. As to whether it will make blockchain obsolete, as some have speculated, that answer is much clearer—blockchain is still king in the land of cryptocurrencies.
