Earlier in the week, I spoke to Xinshu Dong, CEO of the Zilliqa (pronounced Zilli-ka) public blockchain platform. Xinshu and I spoke on a variety of topics — ranging from his foray into the blockchain world, how Zilliqa is addressing the enterprise scalability needs and the future roadmap.
Excerpts from the interview…
Xinshu, how did you journey into the blockchain space?
My journey into blockchain started with research at the National University of Singapore. Around the time of research, I got an opportunity to work on a blockchain project with the Singapore Stock Exchange, as the technical lead for the project. I was working on building use cases for the private blockchain network and in the middle of 2017, I recognised the importance of scaling up into the public blockchain space. That’s when I made the switch.
What’s your vision for Zilliqa?
My vision is to make Zilliqa a de-facto platform of choice, for people wanting to build high-throughput, high-scalability applications. The vision is also to develop Zilliqa into a general platform serving cross-industry purposes and needs. To begin with, we are focused on specific sectors, to demonstrate that concrete use cases can be built on Zilliqa.
Moving onto the core of this discussion, scalability is a real concern for enterprises wanting to adopt blockchain. How are you addressing it?
We use many methods to address the scalability challenge, the notable one is the sharding technology.
Sharding is a classical concept and we use transaction sharding at Zilliqa. The core idea is to divide the interconnected nodes (or computers) into smaller groups, and each group then process some of the transactions. Eventually, all the transactions get aggregated. Unlike other sharding techniques (for example the state sharding), in transaction sharding, every node doesn’t need to store everything. Yet, every node in the entire blockchain network is synchronised by sharing the same state. This feature gives the flexibility to increasing processing capacity, by simply adding nodes.
We’ve also redesigned the structure of the blocks — the basic unit for storing information in the blockchain networks. Simplifying the storage and transferring of the blocks, like sending hashes or only the changes, has helped us add to the performance.
Our consensus protocol is based on PBFT (Practical Byzantine Fault Tolerance) and we’ve made improvements to it through multi-signatures (Schnorr signature). Without multi-signatures, when nodes were communicating with each other, it resulted in bulky messages. Using multi-signatures, we’ve been able to control the size of the message. This prevents the network from slowing down.
One common perception is, sharding leads to latency issues. What’s been your experience?
Sharding does not make latency a bigger problem, nor sharding is a solution to latency. It maintains the status quo. To understand latency, we need to look at the fundamentals. Blockchain networks are slower than centralised solutions. This is a feature of the blockchain networks. Blockchain uses redundancy to achieve better resilience, security and decentralisation. The cost of having redundancy is higher latency.
However, keeping the experience of the application users, there are options to improve on the latency. For example, user confirmations can be given much faster (i.e., optimistically giving results to the users with a negligible error margin) compared to the actual latency of the blocks.
Given the variability in achieving scalability, what considerations should the enterprises have in mind before implementing blockchain?
The best approach is to start with a pain point requiring better security, efficiency and management procedure. These pain points may exist across their process, products or protocols. Then, see whether blockchain is needed to solve those pain points.
Certain areas where blockchain can help improve existing conditions are data security to prevent the loss of data, reducing the multi-party collaboration cycles, and using immutability aspects to prevent the dangers of internal or external malicious actors.
In short, blockchain cannot solve all problems. Start with the pain point and look at matching blockchain based solutions to solve them.
Finally, what is your advice to the people on-the-fence regarding blockchain?
Now is the best time for everyone to be humbled to learn. This applies not only to people on-the-fence, but also blockchain advocates. The hype is high, and this is a future technology. The people on-the-fence are also going to be very useful, because they provide the criticism and going to ask critical questions for people like us to address, before they are going to be convinced.
My advice for them is to engage with us and have dialogue with us. Let us know your exact concerns and issues in adopting blockchain technology. This will help us improve the technology to help with their use cases.
Full disclosure: I have no association with Zilliqa at the time of this interview.