From my tests 1.2 million transactions per second in a true decentralized, distributed, on-chain, single shard environment is currently not possible with the hardware that exists today. Nor can it be byzantine fault tolerant. 1.2 million transactions per second on a high end server assuming everything is in memory already and you are only limited by the CPU is theoretically doable, but, you want to start adding disk cache, look at around 100k tps, you want to add network, look around 20k tps, you want to have it byzantine fault tolerant, look significantly lower.
Zilliqa beat Ethereum to the punch in releasing sharding. In Zilliqa’s most recent test run they claim 2.5K TPS with 6 shards, 3,600 nodes on homogenous Amazon AWS EC2 instances. They even claim to have exhausted all of the available AWS instances in Singapore during their test. So what happens when the difficulty level increases and the nodes introduce varying qualities of hardware. Likely, TPS will decline. Zilliqa’s latest results clearly show they are moving along in increasing TPS with an increased number of shards and nodes. Their test 7 months ago hit 1,400 TPS with 4 shards, 2,400 nodes).
So first place for me to go is Quarkchain’s GitHub repo (here).
Ghost town. No public repo. No contributors. You can see the light and dark green boxes in the contributions section. Most of the activity was back in February and April. Basically the code is sitting behind closed doors in a private GitHub repo. Since there’s no code to examine without getting access, let’s dig into the architecture instead.
[4/17/2018 update] Andre Cronje’s CryptoBriefing review came out. Definitely read this immediately after reading my post. He goes in much more depth on the technical aspects of this project. TL;DR… their implementation is novel and well executed.
Parent and Sharded Chains
- Two-tier chain architecture: the first tier is the rootchain and the second is the sharding layer. The rootchain only checks the headers of the sharded level blocks. Each sharded chain houses their own transactions. The sharded chains will feature cross chain communication. Quarkchain has opted for horizontal sharding.
- Network bandwidth: Let’s start with their bold claim of a million TPS. I don’t believe it but let’s just call this their theoretical limit when they fully wrap up their development. – up to a million transactions per second, which is much more than the current capacity of Visa and Mastercard combined. Their testnet run posted on their site shows 2,300 TPS. This is similar to Zilliqa’s test last month (2,500 TPS) with 6 shards, 36 nodes.
- Cross-Chain transactions: Transactions across sharded chains. This is clearly the defining feature of this project. Imagine
- Smart contracts: Par for the course these days. Turing complete smart contracts.
Quarkchain is currently looking for 100 people to stress test their testnet. Here is their announcement looking for experienced blockchain programmers and network engineers.
Qi Zhou, Founder and CEO. Short stints at EMC, Google, and Facebook. To get hired at each of those firms is a testament to his ability. One knock on his experience is how short of a time he spent at each. Those criticisms are less important for entrepreneurs. His experience at each of these firms revolves around sharding and clustering. Those are right skill sets.
Yaodong Yang, University of Michigan (as an alum… Go Blue!). Experience at Google, Instagram/Facebook. Likely knows Qi from Facebook.
Xiaoli Ma Ph.D. Electrical Engineering, University of Minnesota. Current research in signal processing for communications and networks. Useful skillset for the project. She brings strong academic experience as a professor at Georgia Tech and an Institute of Electrical and Electronics Engineers Fellow.
The rest of the team are missing LinkedIn profiles. I’m going to just call those red flags for now.
Arun Phadke – Professor Emeritus of Electrical Engineering at Virginia Tech. Ph.D. from University of Wisconsin. Area of focus is energy engineering.
Leo Wang – Head of PreAngel VC (a large Chinese venture fund focused on early stage startups). Qlink advisor.
Kevin Hsu – Masters of Computational Statistics and Machine Learning from UCL. Partner in Dragonmind Capital. Blockchain investor, CEO/Founder of Block VC Fund.
Zhiyun Qian – Experience at Microsoft and Cisco. Assistant Professor at University of California Riverside. Focused on cyber security.
Quarkchain is trying to solve the “really, really fast blockchain” problem. So are a whole slew of other teams as I mentioned in this post. Zilliqa is testing their solution and have proven their testnet is up and running. Ethereum’s sharding solution is supposedly underway. Credits is in flames in the public eye. With the race to make sharding work, it’s a positive development Quarkchain is actively recruiting testers. Their private repo must at least be at a level worth showcasing to testers. Please apply if you fit the bill. I’d love to see how their technology performs.
I’m not impressed by big headlines such as “million TPS”. It’s clickbait at best. But I am intrigued by the team’s background. There’s a wealth of relevant experience on the core team and the advisory board. Their team is loaded with big name firms (Google, Facebook) and with PhD’s in the relevant fields. If you’re going to bet on a speculative project, you’d want to bet on a team with backgrounds like these. The Team reminds me of the RChain team: very academic, big tech experience.
In addition to their backgrounds, the team is located in blockchain hubs. They are incorporated in Singapore (like every other major blockchain it seems) and many are present in the Silicon Valley. Proximity to so many other projects and advisors helps in intangible ways and directly when it comes to recruiting future team members.
Their GitHub repo is private at the moment but it would be nice if they opened it up once the testnet results are in. It would bring goodwill to the community and recruit others to provide suggestions/improvements to it.
As for the token metrics, it’s impossible to say what we are looking at in terms of price evaluation. Token metrics to be announced. I know that their private sale has already closed. Their public sale is targeted for the end of May or beginning of June. Hard cap of $20M.
Seeing that Zilliqa did well in its ICO (9x USD, 13.7x ETH) I would assume Quarkchain to perform well for its investors.
If you like my content:
- Please consider donating. My blog’s ethereum address:
- Follow me on Twitter (@CryptoQuantalys)
- Follow me on Steem (@quantalysus)
- Coin Review: Ontology
- Coin Review: Aelf
- Coin Review: Banyan Network (BBN)
- Coin Review: Mithril: a social network app on the blockchain
- Coin Review: Qtum
- Coin Review: Waves
- ICO Review: DAOStack
- ICO Review: Alchemint
- ICO Review: Loki Network
- Opinion: Steemit’s Smart Media Tokens
- Opinion: Why we need regulation
- Opinion: Token economics
- Opinion: ICO paradox
- Research: Cost to run an ICO