ZILLIQA vs Cardano
There are three aspects wrt which Cardano and ZILLIQA can be compared:
1. Approach to scalability: The current approach that Cardano takes towards scalability is PoS. ZILLIQA’s approach to scalability is sharding with pBFT as the underlying consensus protocol and PoW as a Sybil resistance mechanism.
PoS alone does not give linear scalability as sharding does. Cardano in its design philosophy1 claims that sharding becomes trivial to implement with PoS. We disagree with this statement as sharding is known to introduce multiple challenges including but not limited to: how to merge transactions processed from different shards, how to safely and efficiently transfer transactions processed in one shard to the other shards without creating a bottleneck, etc.
With smart contracts, sharding introduces further challenges. To see this, consider a smart contract that runs an ICO, i.e., it sells a finite number of ERC20-type tokens in exchange for some native Cardano tokens. Each contributor sends a transaction specifying the required number of Cardano tokens and in return gets some number of ERC20-type tokens depending on the price of the token.
Now consider the case, where, the contract has only 1 remaining token and two contributors say Alice and Bob wanting to buy the last remaining token. In a sharded environment, Alice and Bob’s transactions may get handled in different shards. In the absence of cross-shard communication prior to creating the next block, each shard is oblivious of the fact that the token has been allocated to the other user in some other shard. Such conflicts need to be resolved either synchronously by communicating with other shards or could be settled a posteriori after the two shards have proposed their blocks. Conflict resolution creates challenges that are non-trivial to solve if we want to maintain scalability and avoid computation and bandwidth bottlenecks.
2. Consensus protocol: The consensus protocol in Cardano is PoS, while ZILLIQA uses pBFT. The two consensus protocols work under different threat models. PoS assumes a cryptoeconomic model where the adversary is not willing to lose a collateral for malicious interests. On the other hand, pBFT assumes a harsher adversary model where miners are Byzantine and can arbitrarily deviate from the protocol. The other advantage that pBFT has over PoS is finality — this means that ZILLIQA does not need confirmations.
3. Development principle: Cardano’s most recent release Byron (September, 2017) is an implemen- tation with centralized nodes. Cardano aims to eventually transition towards a decentralized architecture. ZILLIQA’s development philosophy is different. We believe that such transitions are prone to issues and may require an enormous change in the codebase that may take a considerable amount of time. With ZILLIQA, we have built a decentralized infrastructure from the ground up and hence can potentially provide an accelerated path to a scalable solution.