Zilliqa — A Sharding Solution from Academia
TL;DR. Ecology:3. Technology:6. Decentralization:5. Valuation:5. Rating:5/10
(Update 2019-01-09: Please see the semi-annual ratings review of ZIL here.)
Zilliqa is a project focused on creating a blockchain with extremely high Visa-level transaction throughput capacity. The chosen way to achieve this is by sharding, a method where network nodes are divided in groups that each runs the consensus protocol in a siloed manner. Another dedicated group processes the contributions of all shards and reaches consensus on an aggregate level. Zilliqa calls this method divide and conquer. To simplify the advantages of transaction sharding, imagine that each shard can process X transactions per second. By processing transactions in parallel, Y shards of the same network can process a total of X*Y transactions per second. In other words, the network throughput scales linearly. The Zilliqa team members seems to have been among the first to research sharding for public blockchains. Many of them, including Zilliqa CEO Xinshu Dong, are from the National University of Singapore.
Going a little bit deeper in to the consensus rules mentioned above, BD Ratings concludes that Zilliqa is utilizing a mix between Proof of Work (PoW) and practical Byzantine Fault Tolerance (pBFT). Zilliqa aims at having Sybil resistance through PoW, while still keeping the total computational costs down for the network as a whole by using a pBFT algorithm and not PoW for establishing consensus. The PoW algorithm establishes "mining identities" by forcing new nodes to present enough proven computational work before they are randomly assigned to shards where consensus is reached with pBFT. As any node may submit PoW to the network, there is no way of knowing if it is malicious or not, but it helps reduce the total number of malicious nodes that would otherwise threaten the integrity of the shards.
Zilliqa will support smart contracts with Scilla, a new intermediate-level language. Tools are developed that, if successful, may convert dapps built on Ethereum's popular Solidity language to Zilliqa compatible code. The Zilliqa code base itself is novel, with bits and pieces taken (like the Ethash PoW mining algorithm) from other public blockchains.
A pre-ICO was conducted in the end of 2017, and due to the rising ETH prices back then, the USD 20M hardcap hit before a planned ICO could start. Instead, community members were allowed to contribute some small amounts before the Token Generation Event took place, resulting in another USD 2M funding. 60% of the 21 billion maximum Zilling (ZIL) supply is minted as ERC20 tokens now while mainnet is being built. Half of it went to ICO contributors (and supporters), while the other half went to the Zilliqa team, their company, and some external agencies; most of it is vested and continuously released over a 3-year period. The last 40% of all tokens are reserved for mining rewards over a ten year period once mainnet has launched. That minting is for subsidizing miners in the absence of enough on-chain transactions
A public testnet was released in April 2018, and the second version of it was just released in July the same year. With the latest release, Scilla templates like a fungible token contract and a crowdfunding contract were released as well. On a private testnet, the Zilliqa team managed to have the chain process around 2000 transactions per second. With a larger pool of nodes, the team thinks 10 000 transactions per second could be reached.
ZIL tokens exists only as the Ethereum ERC20 standard for now, so it is not simple to correctly estimate economic activity as all the token transfers are just for the purpose of speculative trading and consequent storage. As soon as mainnet is live (most likely in 2018) and there are a couple of dapps up and running, real interest can be easier measured and compared to other similar chains.
The official subreddit is active but not precisely bursting with activity. The Zilliqa team is continuously communicating with the community on Reddit and through blog posts. A USD 5M grant programme has recently been introduced as well, which ought to expand the ecosystem. BD Ratings have a feeling that community activity is quite thin for many cryptocurrencies right now as prices generally have slumped more than 60-70%.
Although conversion tools might help steer smart contracts from Ethereum to Zilliqa, the fact that Scilla is a lower-level language could act as a threshold to projects in search for a blockchain. On the other hand, one could argue that projects prioritizing easy development might not belong on an open blockchain anyway because of the high security risk in case of bugs in the smart contract code. This balance is still somewhat diffuse but has shifted towards the security part of the spectrum after serious bugs in Ethereum contracts affected many.
Finally, Zilliqa seems to divert a bit from many other cryptocurrencies in that they are not spending a lot of time promoting their token. BD Ratings thinks that this strategy, where marketing and other "visuals" are downplayed and focus rather is on GitHub repositories, will win in the end. It creates a solid, loyal community rather than a larger bunch of bounty hunters.
Reasons: Quite low community activity. No mainnet yet.
Technology seems to be the main strength of Zilliqa. The project has its roots in academia, as many on the developer-heavy team were among the researchers that first published findings on sharding in public blockchains. In a reddit thread, user Coor_123 summarizes the academic contributions from many Zilliqa team members. Xinshu Dong, Zilliqa CEO, have a multitude of cited publications in his baggage, which seems to be focused mainly on cyber security and its implications for public infrastructure. Yaoqi Jia, co-founder and Head of Technology, has a Ph.D in Computer Security from National University of Singapore. Prateek Saxena is advising the project. He is an Assistant Professor at the same university and also a co-founder of Anquan, the research company with deep ties to Zilliqa.
With the academic background, a broad team of developers, an active GitHub, a working testnet and a first mover advantage on a sharding scaling solution, Zilliqa has a strong Technology rating overall. The team focuses on the technology rather than having a marketing approach. Its sharding technology is discussed briefly in the Ethereum Sharding FAQ, where a "critique" to Zilliqa's transaction sharding is that it increases efficiency only to some degree as nodes still have to store the full blockchain state and thus a large amount of data. For Zilliqa to fully succeed, the Ethereum Sharding FAQ argues, state sharding is necessary as well to curb centralization. In this Technology section however, Zilliqa's transaction sharding is a very positive parameter and will lead to an extreme increase in transaction throughput. Congestion as seen last year in the Bitcoin and Ethereum blockchain will be highly unlikely with Zilliqa's close to 100x optimization.
The novel code base together with the novel pBFT consensus rule is a quality stamp on the whole project. It is common for smaller cryptocurrencies to simply copy the code base of a larger battle tested coins, but as Zilliqa has a clear scaling focus, they have chosen to build from the ground up with this goal in mind. Untested clients and algorithms may very well carry a higher risk of failure, but on the whole, it is in this case a positive thing for the Technology rating.
One claim that BD Ratings was not able to fully verify is the Zilliqa claim of lower electricity footprint due to the PoW-pBFT combined consensus algorithms. If passing the PoW participation threshold entitles a miner to high (non computational-intensive) pBFT block rewards, then they will surely point a lot of hash power at Zilliqa during the short PoW period, making the average hash rate rise towards that of a fully PoW benchmark like Ethereum. The Zilliqa team has not expanded on this issue and may overstate the reduction in electricity consumption.
Reasons: Strong academic team. Active GitHub. No mainnet but working testnet. Sharding technology.
Even though the team behind Zilliqa seems to have a good understanding of the importance of decentralization for a public blockchain, there are a number of issues that ought to be explored. In a recent AMA on Reddit, the Zilliqa team acknowledges that because of the high transaction throughput, any full node would need to store terabytes of data if the chain is indeed used to capacity. Because of this, Zilliqa is in touch with storage network platforms like Bluzelle and Genaro to see if these could provide acceptable solutions. This may create single point of failures where these companies could be pressured to stop hosting the Zilliqa blockchain data, even though they seem to aim for decentralization as well.
There is more evidence of Zilliqa focusing on the decentralization principle. Rather than building first, then (hopefully) entering a decentralization process like Ripple or NEO, Zilliqa has built the technology to be decentralized from the start. By passing the PoW threshold, anyone can participate in the pBFT block creating process. Also, the developer team is large enough to not create obvious dependencies, and the open source code means that anyone can build on the project.
As token ownership is not affecting the chance of obtaining block rewards (as is the case in many PoS chains) the ~27% of all ZIL supply that will end up in the hands of Zilliqa, its founders or Anquan will have no huge centralization effect on the network.
Reasons: Blockchain storage issues. Decentralization principles evident withing the project.
Given that almost all the ZIL supply will hit the market within the coming years, BD Ratings chose to include all 12.6B tokens in the market valuation. That leaves us with a Zilliqa valuation of around USD 1B, which is quite high given that the market as a whole has crashed. This should however be seen in relation to some important properties of Zilliqa: a transaction sharding solution, a decentralization focus, a novel code base, and a technically strong developer team. Mainnet is close enough for BD Ratings not to make the fact that the chain does not yet exist a large discounting factor.
Reasons: Strong team from academia. Sharding solution.