Thunder Token

Pros

  • Experienced team members
  • Interesting approach to solving the performance boundaries of blockchains

Cons

  • So far only 5 team members and the project is in a phase of hiring new talents

Team: 4/5
Project: 5/5
Token use case: ?/5
Token Metrics: ?/4
Roadmap: 1/3
Community and marketing: 2/5
Hype: 1/3

Description

Adding a “fast-path” on top of a standard blockchain called the “slow-chain”. In normal conditions, all transactions are instantly confirmed on the fast-path, and we barely have to use the slow-chain. If the fast-path is ever disrupted, a provably correct mechanism is invoked to fall back to the slow-chain, and nodes can continue to communicate and confirm transactions on the slow-chain. Once the problems are diagnosed, the fast-path will be revived.

 

Team

The team currently consists of 5 people + one intern and as they state on their website, they are hiring engineers, product managers, and others who are committed to creating a high-performance blockchain.

Chris Wang – CEO, Co-Founder

  • VP of Tech in The Walt Disney Company for 2 years
  • Founder of Playdom – Playdom was acquired by Disney Interactive Media Group in 2010.  Playdom produces a diverse portfolio of casual games for the rapidly growing platform of social networks, including Facebook and MySpace. With over 47 million monthly users and #1 rated games.
  • PhD, Computer Science at Carnegie Mellon University

https://www.linkedin.com/in/chris0804/

Elaine Shi – Co-founder and Chief Scientist  

  • Co-founder and co-director of the Initiative for Cryptocurrency and Contracts (IC3)
  • Associate Professor. Computer Science at Cornell University
  • PhD, Computer Science at Carnegie Mellon University

https://twitter.com/elainershi?lang=en

https://angel.co/elaine-shi-1

Rafael Pass – Scientist

  • Associate Professor at Cornell University
  • PhD, Computer Science at MIT

https://www.linkedin.com/in/rafael-pass-7aa7361/

Jianqing Zhang – Core Engineer

  • Senior Software Engineer at Uber for 1 year and 2 months
  • More than 5 years as Security Researcher at Intel Labs
  • Software Engineer at IBM China for 1 year
  • PhD, Computer Science at University of Illinois at Urbana-Champaign

https://www.linkedin.com/in/jianqingzhang/

Chris Li – Engineering Director

Project

State machine replication has been a central abstraction in the 30 years of distributed systems literature. In a state machine replication protocol, a set of nodes seek to agree on an ever-growing log over time. It requires two critical security properties:

  1. consistency, i.e., all honest node’s logs agree with each other although some nodes may progress faster than others
  2. liveness, i.e., transactions received by honest nodes as input get confirmed in all honest node’s logs quickly.

 

The low latency and poor scalability of Nakamoto’s blockchain protocol are typically viewed as the main bottlenecks for Bitcoin as well as other cryptocurrencies. The Thunderella paradigm provides a very practical and simple approach for overcoming these issue. It shows how to build on top of currently running blockchains, to enable “optimistic instant confirmation” of transactions. Additionally, note that in that protocol, players only need to send transactions to the leader, who in turn lead the committee to confirm the transaction. Most notably, the underlying blockchain is essentially only used when something goes wrong, and blocks need not be distributed to the whole network before getting confirmed; thus, Thunderella also solves the scalability issue with Nakamoto’s blockchain protocol. Of course, both of these guarantees are only “optimistic”— but arguably, under normal circumstances, one would expect 3/4 of the players to act honestly, and the leader could be incentivized (paid) to perform its job (and if it doesn’t, will be kicked out). Thus, the belief is that their approach is a practically viable approach for circumventing the main bottlenecks of today’s cryptocurrencies.

 

Thunderella is also a constant factor faster in the fast path than most PBFT- or Paxos-style protocols. PBFT-style protocols typically require multiple rounds of voting even in the normal path (c.f. Thunderella has exactly one) — and the latter rounds are necessary to prepare for the possibility of a view change. Although it is possible to compress the normal path to a single round of voting, this is typically achieved either by sacrificing resilience (e.g., tolerating only one-fifth of the corruptions) or by adding yet another optimistic layer on top of the normal path — thus further complicating the already complex protocol.

 

Classical modeling techniques in the distributed computing and cryptography literature typically treat crashed nodes as corrupt, and thus crashes would count towards the corruption budget. Thunderella considers a permissioned model in which crashed nodes that are not under adversarial control are treated not as corrupt, but rather as honest but “sleepy”. Such a “sleepy model” was first proposed in the recent work [Rafael Pass and Elaine Shi. The sleepy model of consensus. In Asiacrypt, 2017]. In the sleepy model, the set of online and honest nodes in adjacent rounds may be completely disjoint; and nodes can go to sleep and then wake up later.

At a high level, the sleepy consensus protocol is very similar in nature to the original Nakamoto blockchain — while Nakamoto’s blockchain uses proof-of-work to elect random leaders, the sleepy consensus protocol emulates the random leader election mechanism without relying on proof-of-work.

 

Recent work of Rafael Pass and Elaine Shi showed that the following interesting facts:

  • State machine replication is possible in such a permissioned, sleepy model, as long as in every round, the majority of online nodes are honest
  • Interestingly, no known classical consensus protocol can achieve state machine replication in the sleepy model, even when we are guaranteed that 99% of online nodes must be honest in every round.
  • Moreover, the honest majority assumption (among online nodes) turns out to be necessary to realize state machine replication in the sleepy model.

 

The basic Thunderella Protocol approach assumes three logical entities:

  • miners of the underlying blockchain;
  • a leader; and
  • a committee

 

Token use case

Most probably the token/currency will be used as a collateral for leaders and also as a value transfer.

 

Token metrics

Not disclosed yet.

 

Roadmap

Not disclosed yet but the testnet should be released in June of 2018.

 

Community

Only social media where they are currently most active is Twitter https://twitter.com/ThunderToken where they have about 1600 followers. On Linkedin https://www.linkedin.com/company/thunder-token-inc/ they have 256 followers and on Facebook https://www.facebook.com/thundertokens/ just 6.

The team is now more active in participating on various events rather than growing their social media presence.