A core part of Incenti News’s mission is to cover interesting projects in the blockchain/DLT space that for various reasons do not receive sufficient coverage by the major crypto media. One of such projects is, without doubt, Hedera Hashgraph.
General info about the project
Hedera Hashgraph is an ambitious project that was launched in early 2018, aimed at creating a highly scalable public DLT platform allowing to both make payments and run smart contracts, as well as store data on it. What may be important for adoption is that Hedera will support Ethereum’s most popular smart contract language Solidity for smart contracts. Currently, the platform functions in the form of a restricted-access mainnet, and the plans are to gradually open it up together with the issuance of the native HBAR token to the early investors.
The key distinguishing features: the Hashgraph consensus algorithm and the event DAG
By far the most distinguishing feature of Hedera is its Hashgraph consensus algorithm created by the project’s founder computer scientist Leemon Baird coupled with the unique directed acyclic graph (DAG) of events data structure that goes together with it.The communication among nodes is based on a gossip about gossip approach. A good introductory explanation of how the consensus approach works can be found in this presentation by Baird.
As we understand it, instead of a chain of blocks, Hedera is based on an arrangement where a node uses its own latest event and the latest event reported to it by another node to create a further event and sign it with its digital signature.
Events are like blocks in the traditional blockchain context but they may in practice be smaller, and if the node wishes, its event may contain no new transactions and just references to the parent events. In the latter sense, it is similar to the blockchain because the integrity of the data structure is provided by the hash links and electronic signatures of block creators. The evolution of Hedera’s data structure, however, is asynchronous and nodes can add events in parallel without trying to add a block to the highest block in the chain, hence the DAG structure.
As the direction of the DAG is given by time, the DAG is split into time periods. The first event for each node that publishes an event during a period is called ‘witness’. The purported advantage of the Hashgraph algorithm is that it involves only doing computations on witnesses in each period.
In the period after the event in question was submitted by its creator, it needs to be referenced (“seen”) by at least two thirds of the witnesses in one of the following realms. After this if each of those witnesses is seen by at least two thirds of the witnesses in one of the following realms, it means that the original witness becomes “famous.” The performance of this procedure on witnesses that is, according to Baird, usually pretty fast and straightforward, is sufficient to validate the transactions from the same realm as them.
It has to be noted that the actual workings of the mechanism appear to be quite a bit more complicated, as nodes may also have different amounts of the network currency that they stake and so on, the majority is calculated in terms of stake, nodes must take the influence of other nodes based on their relative stake into account and so on.
Claimed advantages of the platform
The team behind the Hedera project has claimed several important selling points for the platform, such as:
- the purportedly proven soundness (more formally, asynchronous Byzantine fault tolerance) of the consensus mechanism, which is based both on mathematics and computer verification using Coq code. It has to be noted, though, that the claims and proofs in question do not appear to have been peer-reviewed but instead merely published as Swirlds (the original company behind Hashgraph) technical reports;
- very high scalability achieved thanks to the DAG structure, asynchronous consensus and sharding;
- genuine, rather than probabilistic, transaction finality. Existing PoW blockchains at least can only provide probabilistic finality. If an attacker gains control over the majority of the nodes, it may, in theory, given enough time, re-do proof-of-work on past transactions and revert those that she wishes, even though, in reality, this is impractical;
- fairness of transactions guaranteed by the fact that the consensus algorithm timestamps them based on the average time nodes became aware of them rather than the time of inclusion into the block as in the traditional blockchain approach;
- low, albeit non-zero, transaction cost that could make on-chain micro-payments feasible, especially given that the form of transaction confirmation can be tailored to the transaction’s importance;
- parallelization of smart contracts using the logical tool of the so-called “realms”;
- on-platform storage of data.
Team and finances
Hedera’s team led by Baird and Mance Harmon appears to be rather solid, even though the project is extremely ambitious, too.
The project has also raised more than $120 million in investment, and, according to Harmon, its treasury is holding them in fiat, which guarantees that the project will have sufficient funds for years of ongoing development.
According to an article on the Hedera’s official Medium page, in September last year, there were already over 100 DApps being built on the platform. The seven most prominent projects mentioned in the article were Sagewise (software to resolve blockchain disputes), Carbon (a stablecoin project), Hearo.fm (a blockchain-based music streaming marketplace), Alto (a platform for issuing and managing non-fungible digital assets), Red Swan (real estate tokenization), Cryptotask (online marketplace for freelancers) and Artbit (micropayments for online art performances).
Network token supply and distribution
According to its CEO, Hedera has already generated all the 50 billion native HBAR tokens that will ever exist, although the vast majority of them are not yet in circulation. A crowd-sale related official post on Medium promised the following projected distribution at the beginning:
- 65%: Hedera Council Treasury
- 17%: Hedera management and employees
- 13%: SAFT purchasers and developers
- 5%: Swirlds
This is, however, subject to change as only 10% of the token supply is projected to be released by the end of the first year of Hedera’s existence. Overall, the fixed, slowly-emerging supply model chosen by Hedera appears to be similar to that of many other projects in the space. In order to address the potential price-volatility issues this approach may cause, Carbon project is developing an algorithmic central bank-style stablecoin on Hedera, although the recent failure of a similar Basis project on Ethereum due to regulatory obstacles makes Carbon’s prospect’s look unclear.
Current state and roadmap
In August last year, Hedera launched a restricted-access mainnet version in order to allow projects and developers to gradually test it with the possibility to earn some hbars in the process. Last December, it was opened to community testing and building third-party applications, allowing the willing community members to register for access on its website. Phase I was apparently focused on micropayments and was completed in the end of January with the creation of 5000 accounts on the platform. The registration for Phase II is now finished but it may be reactivated, according to the web-site announcement. At the same time, it is not wholly clear from the publicly available sources when the open-access version of the mainnet will launch.
Governance and the licensed nature of Hashgraph
Hedera’s approach to protocol governance is perhaps unique in the crypto space. Governance is separated from consensus and is inspired by the original model of the association that later became Visa.
On February 20, the initial members of the Governing Council were named, such as Deutsche Telekom, DLA Piper, Magazine Luiza, Nomura Holdings, Inc., and Swisscom Blockchain AG. In the future, the council is supposed to have 39 members from 18 business sectors and diverse regions. Each of them will have an equal 2.6% stake in the system and will have term limits, with the exception of Swirlds.
Despite the claimed separation of consensus from governance, however, initially, it will be the 39 governing members that will be running the nodes is the first open-access version of the Hedera network, more precisely, the first shard that will allow for up to 10,000 transactions per second.
It is also essential to note with regard to protocol governance that the Hashgraph technology will remain proprietary to Swirlds from the legal standpoint and will only be licensed to the Hedera Council in exchanged for a guaranteed share of revenue.
To our knowledge, two major criticisms have so far been raised about Hedera, one related to its consensus mechanism and one to governance structure. Esteemed cryptocurrency researcher Emin Gun Sirer told Forbes a year ago that he doubted Hedera’s claims about the soundness of its consensus mechanism:
The correctness of the entire HashGraph protocol seems to hinge on every participant knowing and agreeing upon N, the total number of participants in the system. This is a difficult number to determine in an open distributed system.
In response, Baird voiced his disagreement claiming that all the nodes are aware at any moment how many nodes are there overall. Since then, there does not appear to have been any more back and forth between them in this regard.
The second objection raised by many in the blockchain community against Hedera concerns its governance structure, namely, the fact that the protocol will be managed by a group of major companies and that it is not open-source. The sceptics believe that this approach is far removed from the original conception of how a decentralized public ledger should be managed.
Hedera’s supporters’ main response to this claim is to say that even the existing PoW blockchains like Bitcoin and Ethereum (let alone EOS or Tron) are, in fact, rather centralized as far as mining is concerned, and often with regard to the core development activity, too. They maintain that having representatives of prominent companies from different areas of activity and regions in the council will boost trust and ensure long-term stability. They also assert that the proprietary nature of the basic technology ensures that damaging, community-splitting forks are not possible.
Overall, the Hedera project seems to be a solid and promising competitor to both payment and smart contract-focused public blockchain/DLT projects. However, it is too early to say at this stage whether its promises concerning scalability, consensus mechanism functionality and security and so on will be fully delivered upon. It is also unclear to what extent its chosen governance model will allow it to stay a genuinely decentralized problem, at least to the same extent as the major currently functioning relatively decentralized blockchains like Bitcoin and Ethereum.