Aion — An Interoperability Protocol

10.05.2019

TL;DR. Ecology:4. Technology:4. Decentralization:4. Valuation:6. Rating:5/10

Token Sale Dynamics

The first Aion blog post was published on the 5th of September, 2017, and briefly addressed some of the limiting aspects of the cryptocurrency space, like scalability, performance, and interoperability. This new project was initiated by Nuco Inc. - a Canadian company launched some months earlier by three ex Deloitte blockchain leads. Matthew Spoke, Nuco CEO, and his co-founders quickly set out to seek early investment for their venture.

The main Aion project goal was to facilitate an interoperability infrastructure, initially dubbed the Multi-Tier Blockchain Network (MTBN), between public blockchains so that users on one of those blockchains could seamlessly interact with the others in terms of transactions as well as logic. This infrastructure was designed to be both decentralized and incentivized, meaning interoperability dependence on any central entity was meant to be minimized. The Aion public blockchain was dubbed Aion-1. The whole network was later to utilize a Byzantine Fault Tolerant (BFT) algorithm for the bridging protocol, and a PoW-dPoS hybrid mechanism to achieve consensus on the native Aion-1 blockchain itself. In practice, the latter construction seems to have been partially similar to the delegated model explored by the BitShares team and Lisk, meaning throughput could increase by sacrificing the number of validator nodes. But unlike BitShares or Lisk, Aion's goal was to also employ a Proof of Work component called Proof of Intelligence (PoI) in it's consensus mechanism, where validators had to run artificial intelligence computations. The intent was to motivate the creation of AI-specialized hardware that could be used for machine learning and neural network training in the future. Before the PoI algorithm was fully developed, however, Aion was to employ a more conventional PoW algorithm to start with.

On September 24th, 2017, details on the presale and public sale were released by the team. 30M tokens had already been allocated to Aion partners and early supporters. This was done by offering these parties to buy USD 15M worth of tokens for USD 0.5 each. The public presale, with a hardcap of USD 15M as well, was to release between 0 and 40M tokens further, depending on participation rate. The public sale therefore had an initial aim to raise USD 70M, making the full total sum up to USD 100M. As a result of this USD 100M hard cap, it was estimated that around 280M tokens were going to be created in total (30M + 17M +233M) for the public. As 40% of total supply was to be allocated to the founding company and the Aion Foundation (incrementally vested over three years), that caused the total number of tokens to be 465M.

A Token Release Schedule (TRS) contract was created on the Ethereum blockchain, set to activate shortly after the public presale closed. Private sale and public presale participants were eligible to participate in the TRS as a form of incentivized commitment to the project as their tokens would be locked up. Participation would make contributors eligible for a generous bonus in the form of future unsold public sale tokens. All tokens sent to the TRS contract would have a lockup period stretching over a year, where 25% of total tokens were to be returned to the owner as soon as the public sale ended. The other 75% of total tokens were scheduled to be returned in 6.25% batches, released each month over a 12-month period.

As soon as the private sale and presale were over, around 466M AION ERC20 tokens were minted and so the public TRS contract could be used for a subset of these. The public presale never hit the USD 15M hardcap, and so only reserved around 9.5M tokens. Upon the completion of these two phases, the public sale mechanics got updated in the form of splitting it in two time periods- each of which would have 10M AION tokens available. Therefore the total number of tokens that could be bought in the public sale dropped from 70M to 20M, meaning potential TRS contributor bonuses went up drastically instead. This was not the last change in token sale dynamics however. On November 24th, 2017, the team cancelled the two public sale phases altogether, possibly as a result of the public presale not hitting its cap. As the 20M tokens earlier allocated for the public sale now could be partially added to the TRS instead, minimum TRS redistribution bonus increased to 292.73%.

AION Supply mechanics, parallel and cumulative. The small steps represent TRS releases. Taken from a Medium article by Joseph P. DiPasquale. 

Also in November, it was revealed that Aion, ICON and Wanchain together had decided to form the Blockchain Interoperability Alliance - a collaboration which had as a goal to help facilitate a common set of interoperability standards. The logic was that without coordination, there would be a large amount of redundant work and overlap within all these separate projects.

On 8th of December, 217, the public TRS closed with a 91% participation rate. Due to the participation rate not going all the way up to 100%, the effective TRS bonus hit a massive 322.4%.

Kilimanjaro Mainnet

On January 26th, 2018, a schedule for the testnet was announced, and a month later, the whole PoW release of Aion was open sourced. Some months before mainnet, the team introduced FastVM - a virtual machine based on the Ethereum Virtual Machine, meaning it could easily run smart contracts written in for example Solidity.

Aion adopted the Zcash Equihash scheme as initial mining algorithm, but tweaked it somewhat with regards to how much memory was required (and subsequently dubbed it Equihash2109). The rationale for the tweaks was that a multitude of other minor blockchains had adopted Equihash, which tipped the whole algorithm within the profit margin of ASIC manufacturers. Since Aion still aimed at being ASIC resistant, it was deemed necessary to do the tweaks and so partially invalidate the ASICs operating on the Zcash chain.

Mainnet, dubbed 'Kilimanjaro', was to go live on April 25th, 2018, which it subsequently did. It incentivized the validators through a fixed inflationary model that released 1% of total supply as block rewards, meaning around 4.56M tokens for the first year . As the validators kept looking for new blocks on the now functional PoW blockchain, the Aion team could shift focus to the interoperability features that still were not active. The first bridge to deploy was that to the Ethereum network, meaning holders of ERC20 Aion tokens were going to be able to swap to native Aion network tokens through this interoperability infrastructure. An alpha version of this bridge was released in June 2018.

On the 12th of June, 2018, an extensive bounty and grants program was launched. The program focused on rewards for finding technical bugs or vulnerabilities, but also included grants for academic research or the development of various tools. On the 5th of July, 2018, the first version of the Aion Desktop Wallet was released. July continued with the Cayman Island based Aion Foundation launching  - a step in the direction of transitioning operations from Nuco Global to an allegedly more independent entity. This resulted in a corporate restructuring that transferred all intellectual property and current assets of Nuco Global to the Foundation. Furthermore, the 153 million tokens (locked in the 3-year TRS branch) would be transferred to the Aion Foundation as well.

Bridge Construction Begins

With the rise of ASICs in the supposedly 'ASIC resistant' subsection of cryptocurrencies, the Aion team in August 2018 decided to double down on this property and started an open source initiative to further strengthen the resistance. The initiative would conduct research with regards to algorithms but also hardware in order to possibly come up with an in-house hardware solution to future ASIC threats.

On August 16th, 2018, the token swap was finally ready to activate through what the team dubbed a trustless Token Transfer Bridge. This particular bridge would be open between September and November, after which it was going to close indefinitely in preparation for more complete bridging infrastructure and attached tokenomics. To prepare for the Token Transfer Bridge, the Aion network updated to 0.3.0 'Moldoveanu Peak' in August as well. September saw some marketing efforts as the blockchain now was live and soon would have its native token. The Ethereum-to-Aion bridge went live on the 19th of September, 2018, and later closed on the 30th of November that year with a 94% participation rate. November also saw the last distribution schedule release for the public branch of the TRS, which consequently closed.

On November 29th, Aion Java Kernel 0.3.2 'Ama Dablam' was released. December 2018 saw, after a long year with constantly falling cryptocurrency prices, increased concern over the financial state of the Aion Foundation who now acted as main steward over development. Though the USD, BTC and ETH reserves slowly dwindled, the organization still had a steady stream of unused AION coins coming from the TRS.

In a post summarizing the state of the Aion network, the team came back to the question of 'useful' Proof of Work - something that the team hoped to implement through the Proof of Intelligence concept. In this summary, the idea was postponed due to being too complicated. Instead, focus on interoperability proceeded, and in January the team could outline what they called the the Aion Transwarp Conduit (TWC). This infrastructure was compared with TCP/IP in the sense that it aimed to be 'dumb' and not tailor-made to any specific type of data transfer.

March 2019 saw yet another major release, this time of Aion Java Kernel 0.3.3 'Kedar peak'. This was followed by 0.3.4 'Mont Blanc' in April. On the 4th of May, 2019, ARRT (Aion Research and Roadmap Team) published the first draft of Unity  -  a hybrid Pow-PoS consensus mechanism. The team deemed the combination a good balance between increasing economic security with PoS, and the robustness of the battle tested real-world computations of PoW. In this initial spec, no slashing conditions were included.

Ecology

The Aion subreddit has shown considerable activity in the past, but is now a more empty space with around a thread per day. Discussion quality is rather high compared to other projects, and many threads detail the technical aspects of interoperability, often by comparing to other interoperability projects. The Aion team does interact with community members on a regular basis. Aion held the large AIONEX conference in Toronto, Canada on May 2nd, 2018. According to sources, this first Aion developer conference was well attended by as many as 650 people. The Rebuild Conference was held just recently in 2019, attracting many attendants as well. When looking at on-chain transaction activity for the Aion network, this Aion block explorer shows more that 9 000 during the last 24 hours. It is very likely block rewards are part of this number.

The use of bounties and grants will likely continue to improve the ecosystem. Acceptance criteria was on an individual basis and concerned activities such as documentation, GitHub work, finding bugs and vulnerabilities etc. The bounties were in other words not for shallow retweet-and-like nonsense seen in most other projects.

BD Ratings sees one problematic aspect with how the private and public presale were conducted. Cancelling the public sale could on one hand be seen as a professional thing to do; if Nuco had enough funds to get the project going, one could argue raising even more funds from a third sale would just be greedy. On the other hand, and this is rather important, the cancellation opened up to the possibilities of insiders pouring contributions in to the presale instead, as they hypothetically knew their contributions would not be diluted by massive influx of funds in a future public sale. Other participants in the public presale did not know about the cancellation plans, and so had other facts at hand when deciding how much funds to invest. It is worth remembering that the initial plan for the public sale was to raise as much as USD 70M, later changed to a total of USD 20M. Either of those options would have diluted earlier contributions.

As we move from the sales to the actual token swap, there are some problems worth mentioning as well. As has already been established, the token swap window closed on 30th of November, 2018. This is way to close to the last TRS release of 9th of November, that same year. The concern is iterated here. A hypothetical user sending his or her ERC20 tokens to the TRS after the public presale, could be assumed to take note of the last TRS release date. From that data point, he or she might assume that any tokens swap window ought to extend from that last release date and far in to the future. In practice, the Aion team instead set the deadline to only 20 days after the last TRS release. To be fair, the team seems to have had some contingency plans on what to do with tokens that missed this deadline, but on the whole, this token swap event was too rushed.

As the Aion Foundation formed and overtook many stewardship aspects of developing the blockchain as well as managing funds, it is good to see that they set a high standard when it comes to financial transparency. This is probably a result of the EY background of many in the Aion team. The Foundation also conducted public Q&As to clear up any confusion. When looking at the Foundation webpage, it became clear that it is an exempted company incorporated on Cayman Islands. This does not seem very non-profit, and for clarity, BD Ratings reached out regarding this on the Aion subreddit. Almost two days has gone without a proper answer.

But to summarize the Aion ecology: it is a promising project far more professional and serious than most others in the space. As soon as real bridges has been established with accompanying incentive schemes, it will be very interesting to see how value flows between the blockchains in a decentralized way.

Grade: 4

Reasons: Not very high community- and on-chain activity. Very straight forward interoperability use case. Mainnet launched and bridging technology partially tested.

Technology

The Aion team is of considerable size. In 2018, it numbered 60 people, spread out over multiple offices. GitHub commit activity has been high throughout all of 2018, but has decreased somewhat in 2019. All this of course come with the obvious caveat that the size of the team and the number of commits don't immediately translates to real development activity, but it is still a crude but good metric. The earlier mentioned bounties and grants program is likely a net positive for the security of the Aion network.

Well-known Bitcoin developer SDLerner praised the way the Aion team had managed to solve the many security concerns that came with the development of the Java-based Aion Virtual Machine. It is one testimony that the Aion team might not have had to sacrifice smart contract security to reach the world's millions of Java programmers. The team also showed a security focus in how it tried to mitigate the risk of pre-image attacks by letting addresses and public keys be one and the same.

Related to the large size of the team is the funding that the project has, in the form of USD, BTC, ETH and AION reserves of the Aion Foundation. In the latest quarterly report, the Foundation had funds that could last for at least two years, which ought to be enough to finish many of the technical milestones like the Aion Virtual Machine, the hybrid mining/staking algorithm, generalized bridging code etc.

BD Ratings takes issue with the proposed future inclusion of a Proof of Intelligence component in the consensus mechanism. It is admirable that the Aion team wants to make the mining work useful by letting it help 'train' neural networks, but for a public blockchain, security concerns must be of higher importance. Even simple, reiterated SHA256 hashes could be gamed through AsicBoost in the way they were used in Bitcoin's PoW-algorithm. It is then not a large leap to imagine that a more complicated PoW-algorithm could be gamed as well. The proposal of Unity is a step in the right direction; an attacker must obtain more than 50% of total hashrate, and buy up a considerable amount of tokens to stake as well. The solution seems simple enough, and is probably much more secure than any PoI implementation. It addresses conventional double spend attacks, nothing-at-stake attacks, long-range attacks and others. Gridcoin is a protocol where certain miners managed to fake useful work and thus circumvent the utility aspect while pocketing block rewards.

Only with time can we see how the Aion technical specification evolves. There are still many uncertainties.

Grade: 4

Reasons: Mainnet launched. Active development by a considerable team. PoW-PoS hybrid consensus mechanism can probably be highly secure. Technical bounties. Possible negative PoI inclusion.

Decentralization

First, let's explore token allocation within the network. It is important for the Aion network's decentralization property to have as dispersed token ownership map as possible due tot the fact that it partially will run on Proof of Stake. Under the Unity consensus mechanism, owning a token is power, and so should any entity obtain too many tokens, it may start to assert power over the blockchain and lower its integrity. The worst case would be censoring addresses, conducting double spend attacks and so on. So it was troubling when Nuco Global and the Foundation owned 40% of all Aion ERC20 tokens after the October 11, 2017 distribution event. Even more troubling is the fact that this concentration will increase with time as tokens currently vested in a three-year TRS contract is released with bonus up until November 2020. The skewing of token concentration might have partially been an effect of the cancelled public sale (which would have diluted Nuco Global's share of total). In any case, the reality is that it centralizes Aion and would be really, really bad if Unity didn't include a PoW-component.

BD Ratings wants to also address the possible (yet now improbable) use of dPoS (delegated Proof of Stake) as part of the consensus mechanism. The Aion community have had many concerns regarding this issue as well. First example of clear collusion has been seen occurring on the Lisk blockchain, and EOS has had similar experiences. It is probably a net positive that the research resulting in Unity landed on PoS and not dPoS. Aion mining pool hashrate distribution is currently not in a very good shape, with aionpool.tech finding half of all blocks. It was even worse historically. By implementing PoS instead of dPoS in its hybrid consensus mechanism, Aion can hopefully encourage enough stakers to solo stake.

As mentioned earlier, it is not fully clear exactly what kind of legal entity the Aion Foundation really is. In any case, it is unlikely that former Nuco executives would use it to pocket money, mainly because of the high level of transparency in the Foundation financial reports reduces such risk. The Foundation is a net positive for Aion's decentralization rating.

Further development in the right direction is the implementation of a BIP/EIP-like AIP process. By formalizing code change proposals in this way, it reduces the risk that a small number of people can make large changes to the protocol without stakeholders noticing. All AIP's are open for discussion and anyone can read and comment on them.

Lastly, the attempted ASIC resistance is a net positive for the Decentralization rating, but it remains to be seen if the project can keep up with the hardware producers. Monero already has large troubles with keeping it's consensus mechanism ASIC resistant, and they have had to hardfork semi-annually. Unity's PoS-component will help with this kind of resistance.

Grade: 4

Reasons: Token concentration. Hashrate concentration. Unity would decentralize the protocol to a considerable degree.

Valuation

At time of writing, all circulating Aion coins have a total valuation of around USD 57M, which places Aion at #90 on CMC. Counting coming supply from the still active TRS, Aion valuation climbs to around USD 70-80M, or less than 0.1% of BTC. Given the fundamentals, this is somewhat undervalued.

Meanwhile, Cosmos fetches a massive USD ~800M valuation, while ARK valuation is closer to that of Aion's. Hopefully the Aion team will have a hard time increasing the annual mining subsidy of 1% of total supply  -  a number attractive enough to keep the protocol competitive in facilitating value storage. The more this number is finalized, and the more decentralized the project becomes, the 'harder' Aion will be.

Still, BD Ratings sees many unknowns in Aion, both with regards to future PoW-algorithm, ASIC resistance implications, the PoI-algorithm, the deployment of full fledged blockchain bridges and accompanying incentive schemes, blockchain engineer funding etc. Had the price of Aion been lower, a conservative rating would not recommend anyone to buy Aion tokens.

Grade: 6

Reasons: One of the few serious interoperability projects. Ok, but uncertain fundamentals that hopefully can become clearer as the project matures.