The QuarkChain TPS Competition has come to an end completely, and the excellent performance of the participants, which their TPS were over official data, showed us the great potential of the QuarkChain community. We would like to know their thinking about blockchain technology, and many users in the community would like to know more details about the winners. For these reasons, we contacted the first, second, and third prizes of the competition and interviewed them respectively.
Today we first released the third place’s interview, the third winner is a Russian backend engineer. The following is the interview:
1. Could you briefly introduce yourself and your team?
My name is Nikolai Perevozchikov and I’m a Sr Backend Engineer from Russia currently working at gamedev company, blockchain is my hobby.
2. How did you know our TPS Competition?
My brother sent me a link to Medium article introducing the competition after he found it in some Telegram chat group.
3. Why you choose QuarkChain? What other blockchain companies you’re paying attention to except QuarkChain? Compare with other public chains, what strengths you think we QuarkChain have?
I think QuarkChain is great because of it’s sharding capabilities and it’s his great strength. Other companies I pay attention to are: NKN, Ontology, Perlin.
4. You are the 3rd place winner, could you share what did to achieve this high TPS?
QuarkChain ability to host slaves of a single cluster on different machines helped to achieve that result.
5. If there’s no time limitation, what else you think you will do to improve your TPS?
Basically, I think even 100k or more could be achieved easily as long as you have enough powerful machines. Unfortunately, I have instance limits at AWS that didn’t allow me to push my results further.
6. What did you learn from our TPS Competition? For the participant’s perspective, what do you think that the organizer(QuarkChain) can improve in the future?
I learned more about sharding in blockchain. To increase participation rate I think docker image at DockerHub plus instructions for popular cloud providers could’ve helped. Also I asked many questions to Zhaoguang Wang in Discord that greatly helped me to run distributed cluster, but it would’ve been better if that information was shared in documentation for everyone.
7. Thanks for your participation, we believe that you have a deeper understanding of our open source code after joining the competition. Could you share your opinion of our code and any suggestions? Also, as an open source project, we encourage you to contribute more to our code in the future, will you?
I wouldn’t say my understanding is deep, but from what I understand there are a few suggestions:
7.1. I believe currently each slave should connect to every other slave in the same cluster and in my opinion is a bottleneck that won’t allow cluster to scale to more than N slaves.
7.2. I think memory usage could be optimized more, currently it seems too high.
7.3. I believe every cross-shard tx should pass through root shard, which makes root shard bottleneck of the whole system. I don’t know exactly how to achieve that or if that’s even possible, but it would be great if cross-shard transactions could be done between shards directly.
7.4. It would be great to have ability to reshard to custom shard configuration (not just splitting them in 2). Also it would be great to reshard in runtime, without hard fork or even stopping nodes. If QuarkChain in future will have some developer/bug bounty programs I would be happy to participate.
8. How you think about the blockchain industry? What you’re expecting?
I expect rise of many dApps which wouldn’t require to trust to some centralized entity, would be stable (because no company can “shutdown” public blockchain), will make us owners of our own data and could be used seamlessly (like there is no blockchain underneath).
9. Lastly, what you’re expecting for QuarkChain?
For the rise of dApps some problems should be solved, one of greater of that problems is scalability, I expect QuarkChain to greatly contribute to solving that problem (if not solve it completely).
The above is the third place’s interview. If you have any questions, please leave comments. Thank you for supporting QuarkChain. Thank you Mrs. Nikolai Perevozchikov accepted our interview.
This article also published at Medium. Read at Medium: TPS Competition: an Exclusive Third Place Interview