The Zilliqa project was founded in 2017 by a group of researchers from the National University of Singapore. Initial team members included Xinshu Dong, Yaoqi Jia, Amrit Kumar, and Prateek Saxena (who previously worked with Loi Luu, CEO and Co-Founder of Kyber Network), among others. They incorporated the company behind Zilliqa (called Zilliqa Research) in Jun. 2017, a few months before the team launched its first testnet.
Zilliqa brings the theory of sharding to practice with a protocol that aims to address the scalability limitations seen in some other blockchains. Its sharded design allows concurrent chains to execute transactions in parallel, thereby increasing the overall capacity of the network. The team’s primary motivation in building this product was to create a platform that could process smart contracts at scale without sacrificing core blockchain principles, such as the decentralization of node management.
Zilliqa uses a modified version of the Proof-of-Work (PoW) consensus protocol to order and validate transactions, as well as provide security assurances for completed transactions. According to the team, this modified consensus model provides a more cost-effective version of mining when compared to other networks in the market.
Zilliqa's smart contract layer features a proprietary programming language called Scilla, which allows developers to write and express the conditions of smart contracts conveniently. This infrastructure is being used to develop both pilot and commercial applications in the areas of security token offerings, gaming, domain registry service, financial services, and payment reconciliation in advertising.
Zilliqa focuses on scalability through a process called sharding. While many current blockchain applications utilize every node on a network to process every transaction, Zilliqa divides the network into different shards, each processing a subset of the network’s transactions and requiring only some miners working on those transactions at any given time. This reduces wasted computational resources related to nodes confirming transactions that may have already been verified by thousands of others.
Sharding is also applied to smart contract processing, in that only subsets of the entire network resources are used to perform any given computation. Using this technology Zilliqa expects their throughput to roughly double with the addition of every couple hundred miners.
Zilliqa uses standard Proof-of-Work (PoW) for security in its network, but it has a two-layer blockchain structure to allow for more efficient use of computing resources. To create an identity or participate in sharding on a user must show proof-of-work but this is not used to achieve consensus. Instead, Zilliqa employs a consensus protocol known as practical Byzantine fault tolerance. In this protocol, nodes are grouped with a leader that is responsible for broadcasting both transactions and decisions. When two-thirds of nodes in a shard cast their votes, the leader commits an answer. Leader nodes are changed continuously to prevent malicious leaders from having an undue effect on the network.
Unlike Ethereum, Zilliqa does not have a Turing-complete smart contract language. Their smart contract language Scilla was designed with safety in mind and is amenable to formal verification. Zilliqa cites neural nets as being a prime example of a computation that would be best served by their language, as on any other network it would likely be prohibitively expensive.
In the future, Zilliqa plans to research on a significant number of new technologies including:
State sharding – Allowing the data stored by the network to be divided in the same way as the processing and consensus layers.
Secure Proof-of-Stake (SPoS) – A security mechanism that distributes consensus power to the most heavily invested nodes in the network.
Storage pruning – A mechanism for reducing the total amount of data required to store in order to participate in the network. This can eliminate spam and dated blocks from the required data.
Privacy-preserving computation – Computation that removes the identity requirement of participating in the network, so computation is not tied to an account on the blockchain