[Update 5/6/2018], I've spoken with a team member and many of my original, open ended questions have been answered. Please see my original post to compare changes. I did not issue my own verdict but this time around I've edited it to now have a verdict.
What is Edenchain?
Edenchain calls itself a programmable economy platform - the next-generation blockchain platform for a decentralized world. So basically it's a Gen 3 blockchain with smart contract capability. The baseline protocol space is getting crowded as I predicted in my previous post The Multi-Protocol World. My opinion is we are far from moving on to killer onchain applications because no project has captured all three vertices of the DCS triangle (also popularly referred to as the Scalability Trilemma). Secure, decentralized blockchains like Ethereum have not yet solved the scalability problem. Several projects such as Zilliqa and Quarkchain are introducing interesting solutions such as sharding and parallel chains to achieve scalability.
Lately all of my reviews have revolved around Gen 3 platforms. What does this mean exactly? Let's call Bitcoin Gen 1. A blockchain / distributed ledger that incentivized network participants to validate transactions and secure the network. I'm simplifying of course for brevity. Ethereum introduced Gen 2 by showing the possibilities of creating computation capabilities, through smart contracts, within the blockchain. Gen 3 is trying to tackle the obstacles Ethereum is facing: low throughput (as measured by TPS), while maintaining censorship resistance (via decentralization or difficulty), and expensivecomputation and storage onchain.
So let's dive in and see what Edenchain does differently. Before we dive into the architecture let's go through the standard token metrics and team backgrounds.
James Ahn, CEO. Master’s degree in CS from Korea University and KAIST. 20 years of professiona experience. Authored several books and publications in the fields of machine learning and finance. He has presented at technology conferences including Pycon and Open Technet.
Jenny Ryoo, co-founder, Bachelor's from Korea Unviserity and Masters' degrees from Dongguk University. Ten years professional experience. Strong background in cloud analytics.
Kangho Kim, Chief Strategy Officer. Also a Partner at Byzantine Partners (colleague Aaron Tay discussed below is also an advisor). Fixed income and Forex experience at Franklin Templeton.
Team has a heavy presence in Korea but does have boots on the ground in other parts of the world. The LinkedIn profiles below look created for the purpose of Edenchain. It's likely LinkedIn is not a huge platform in Korea - hence why these profiles are so sparse.
Incorporated in Singapore, further legitimizing Singapore as the entity domicile of choice for blockchain projects in Asia.
Advisors
Changki Park, founder of Paxnet and 14 years of experience at Samsung.
Mirza Uddin, Investor at Two Sigma - a NYC based VC focused on early stage consumer, SaaS, Machine Learning, AI, and Blockchain
Aaron Tay, Byzantine Partners, a five person team focused on advisory and investments in the blockchain space - based out of Singapore
Joongheon Kim, Masters and PhD from University of Southern California. Assistant Professor with Chung-Ang University - School of Software (Computer Science andEngineering), Seoul, Republic of Korea, since March 2016; and Director of Distributed Platforms andSecurity Lab at Chung-Ang University.
Edenchain wrote a blog post digging deeper into their team (see here).
A comparison of TPS claims across known protocols and apps
What is their approach?
"According to Ehrsam, 2017, Facebook can handle 157,000 requests per second, and Ethereum can handle 13 transactions per second, and 7 transactions per second in the case of a smart contracted token. In addition, performance and scalability are said to be the most important issues for the blockchain industry."
~Whitepaper page 12
Edenchain attempts to create a blockchain with higher throughput, faster transaction times, and increased security than popular protocols such as Ethereum. In order to see how they plan to do this let's break down their infrastructure. Edenchain will be composed of several layers:
Distributed ledger layer - where data used in the blockchain are separately stored. Different from Ethereum where data is stored onchain. Why is that important? Storage costs offchain are significantly cheaper.
Validation layer - Data transactions are executed and verified here. The EVM (Ethereum's Virtual Machine, explained below) runs the smart contract. To ensure high performance, a transaction scheduler will be implemented. Edenchain's transaction scheduler is based off of Kubernetes, an open-source system for automating deployment, scaling and management of containerized applications. Kubernetes was designed by Google and now maintained by the Cloud Native Computing Foundation1.
Bridge layer - Used to securely import data needed by an on-chain smart contract within the blockchain in cooperation with an off-chain. In the bridge layer, nodes naturally exist on-chain and offchain, and an E-Protocol using ECC-TC, which is an encryption technique, is used for reliable communication between these nodes. (direct citation from the whitepaper)
Edenchain is a permissioned blockchain
Take note because this is fundamentally different from Ethereum and Bitcoin. On both those networks, anyone can participate in the network. On Edenchain, network participants will need to receive access from a network administrator. Bitcoin maximalists and decentralization hawks will decry this architecture. It's against the very ethos of censorship resistance. These camps will even say that this is not blockchain since you have to trust the network validators. I do not take sides here but I wanted to bring it up for you to further study both sides of the argument. So why would Edenchain go with a permissioned network? Basically it comes down to speed. To avoid Sybil attacks on permissionless blockchains, a fee is charged to would-be spammers. With a permissioned network, smart contracts will not need to be stored on every single node in the network. It will also lead to fewer incidents of moral hazard where Ethereum miners will execute smart contracts with the highest gas fees. Edenchain's permissioned state allows it to optimize which contracts are to be run and by which nodes.
Fun facts: other permissioned blockchains include Kadena, Tendermint, Hyperledger
Digging in deeper
Through this architecture it will enable a variety of benefits.
Ability to interface with external data. EdenChain will use what it calls their "E-Bridge layer" to retrieve data when a smart contract is interacted with. The bridge will then encrypt that data with a median voter theorem (MVT) algorithm. This approach differs from Ethereum where computation and storage are run onchain. Ethereum's approach reduces throughput while also serving a very expensive storage solution.
Connecting offchain data to onchain computation is called an Oracle. Edenchain calls their Oracle, "E-Oracle". Original, I hear you. It consists of an E-Oracle client and an E-Oracle server. When the E-Oracle client requests external data, the E-Oracle server collects the data and sends it to the E-Oracle client.
Uses Ethereum's Virtual Machine. Most smart contracts today are written on Ethereum and therefore Solidity is currently the dominant smart contract programming language. Seeding Edenchain's ecosystem will be easier to onramp projects by using Solidity. This is very similar to Qtum's approach (see my past article on Qtum).
Consensus algorithm is called Proof-of-Elapsed Time (PoET). Right off the bat, I like it because it's punny. PoET was first invented by Intel and is currently in use in HyperLedger Sawtooth3. Let's compare the original consensus algorithm in Bitcoin. In BTC, the first to solve the cryptographic puzzle becomes the leader of the network, validating the block. In Byzatine Fault Tolerance algorithms, a series of decisions are made by the nodes. Leadership is chosen in stage. PoET is similar to Raven Coin's X16R and Dfinity's Threshold Relay algorithms. In PoET, the leader is chosen at random by issuing two edicts: 1) generate a random wait time, 2) have the next elected node wait that amount of time. The benefit to PoET is in its potential in reducing energy costs since candidate nodes in the election must power down and wait, while in Bitcoin candidate nodes are constantly competing to solve the Proof-of-Work algorithm.
Concerns
Edenchain's transaction scheduler for its validation layer is built on Kubernetes, an open source project that simplifies building apps on clusters of containers, including those using the popular Docker format2. There's a degree of centralization here since most applications build on Kubernetes are typically housed in one data center. While this ensures high consistency and low partition (see CAP Theorem), it's clearly a tradeoff going for scale rather than decentralization. I mark this as a concern not because it is centralized, but because it potentially serves as a point of failure if the network should be attacked.
If Edenchain is permissioned, can we really call it blockchain? Or is a hybrid solution of permissioned and permissionless states the direction crypto projects are heading in?
Does Edenchain have plans for smart contract development outside of Solidity?
External data processed by the E-Oracle apply an MVT. Basically majority rules. The downside of this is a 51% attack. What happens when a majority of the nodes cannot be trusted? In Bitcoin, a 51% attack is fended off by economic self-interest. In Edenchain's whitepaper their response to this attack is: "In order to prevent such a risk, an E-Oracle consensus module runs programs in the SGX enclave to defend against hardware and software attacks and access external services using HTTPS. Data sources that do not support HTTPS do not allow data the E-Oracle server access."
After speaking with a team member, his response made much more sense. I'll post it down below in quotes but this clearly an area in the whitepaper that could be clarified because it reads as if HTTPS is a defense - which it is clearly not.
...To clarify, HTTPS only ensures that the data is not intercepted when it is being transferred from the external data source into the e-oracle server. MVT is employed to ensure data integrity because each node will ping the data server for information, and if some of these nodes are compromised/data has been tampered with during transmission, the data can still be verified unless there is a 51% attack.
I do not think that the economic incentive is sufficient to fend off a 51% attack in bitcoin. If you look at it from another perspective, other competitors such as Bitcoin Cash has the economic incentive to attack and destabalise the network for bitcoin. This makes it even more dangeous.
In Edenchain, nodes have to stake their tokens to operate the nodes (masternodes), and there is no incentive for node operators to turn rogue at the risk of losing their tokens.
~Edenchain team
Why I like it
The team has the right set of skills.
I'm a fan of random leader selection consensus algorithms.
Offchain data storage should provide cost benefits to applications.
I'm not exactly gung-ho about permissioned blockchains but if the tradeoff is speed for trust, I could see certain use cases willingly make that tradeoff. Just so you know, I don't believe Store of Value (wealth savings) is one of them.
Use of PoET as their consensus algorithm. This represents a potential energy savings for the entire network. The primary criticism will be of a security tradeoff, but security on Edenchain is achieved not through its consensus algorithm but through its permissioned network - only vested parties will host nodes.
Businesses may view this network as a stepping stone into blockchain. Seeing that only permissioned applications can take part may be comforting to IT and Product departments.
Edenchain forked their code from HyperLedger Sawtooth. With that, it inherited PoET. Eden has modified the original codebase. Standard off-the-shelf HyperLedger Sawtooh achieves 40 - 80 TPS (remember, Ehereum is at 15 TPS). Edenchain is targeting 1,000 TPS per namespace.
Outlook: Positive
After re-reviewing Edenchain I've come to view the project positively. Edenchain is approaching the Gen 3 blockchain competition with a hybrid public-private blockchain solution that should be friendlier for businesses to adopt. While businesses will likely consider centralized databases, there will be use cases where a "sufficient" level of decentralization makes sense. Other popular projects such as EOS make a similar tradeoff: trade a little decentralization away in order to gain in speed. For businesses studying the blockchain, speed is preferred over complete decentralization.
Edenchain has a solid code base (Cryptonote/Bytecoin) to work from. They have an impressive team with strong academic and professional backgrounds. They're in the heart of Blockchain-topia in Singapore. I am very interested to hear their game plan to recruit companies to build on their platform. As an analyst, I'd like to size up how much cost savings a business could incur by building on a hybrid public/private blockchain such as Edenchain over traditional database providers or cloud service titans such as Amazon AWS.
With that said, I look forward to personally following them on their success. I wish the team the very best.
P.S. I welcomed the team for additional commentary and they were professional in their reply to me. Professionalism in the crypto space is sorely needed and Edenchain provided it.
Disclaimer: This is not a sponsored review. I am however considering investing in it myself. This writeup is part of my due diligence process and I am sharing it with my followers.
