Introduction

Hi guys, I am back again with another Seele article because some of my readers want more information on their neural consensus algorithm. I would advise you all to read my previous articles on Seele before reading this article, which explain exactly what Seele is. To recap it is a “up scalable neural consensus protocol for high throughput concurrency among large scale heterogeneous nodes with the ability to create a unique heterogeneous forest multi-chain ecosystem” (Seele, 2018).

I will endeavour to dig into the technicalities of this amazing project now and explain exactly how their consensus works, I will also touch on how the network itself operates. This article aims to educate the reader on what Consensus actually is and how Seele’s compares to the other popular ones out there.

As you all know I am extremely bullish on Seele, I see it as one of the all-star blockchain projects available today. Interestingly the name “Seele” comes from the German vocabulary and in English translates as “Soul”; as in the human soul. Effectively I believe their vision is to allow users to create their own vision by using their blockchain, essentially bringing their own value to the internet of today. I truly believe this will be a successful global network for transfer of value.

Seele’s Consensus

In my view Seele and their claim to be a blockchain 4.0 is justified, as you can see from the chart below their consensus and transaction confirmation times are faster than 3.0 projects, they utilise cross chain technology and also have higher concurrency than their competition. All in all, they have far better governance mechanisms than their competition. Seele is pushing the boundaries of blockchain architecture and in my opinion in a league of their own. They achieve this brilliance by ensuring their “node agreement” i.e. Consensus is superior to their competition.

Source: (Seele, 2018)

What is consensus

As I have discussed many times in my articles consensus is the core to any blockchain. Every blockchain has a great consensus driving it from within. In my opinion Seele are far superior to their competition because they use their “neural consensus algorithm” over the others out there. To understand this, we first must look at the basics of what consensus actually is. Consensus algorithms used today by many 3.0 projects do not balance scalability, security and decentralisation very well. Witherspoon, (2018) adds that they exist to prevent double spending on blockchains.

I have described consensus before as the backbone to any blockchain. A simple example of consensus is: If I ask you for 5 dollars, I am conveying a desire to you, and both the receiver and sender agree to this. If either party does not know the meaning behind the request (e.g. language barrier) a problem arises. Consensus involves all parties of the transaction but also the rest of the social group (network) as every entity must consider it a correct transaction for it to be validated (Cryptoguru1, 2018). As I covered in a previous article (which you should read), the main consensuses used today by many blockchain 1.0-3.0 projects are discussed below.

Seele’s main competition

Proof of Work (POW)

This is essentially the first consensus used in blockchain. Bitcoin (Blockchain 1.0) works via this and also Ethereum (Blockchain 2.0), Litecoin and dogecoin (yes, I know it’s funny) are other popular projects utilising this.

This works via computational difficulty so if there are conflicts on the network the hardest chain to generate is kept. In my view this consensus is not economical anymore as the work involved is immense. Castor, (2018) describes it as ensuring the next block is always the “true” and correct version of the truth- which prevents hackers forking and influencing it.

With this blockchain miners compete to add new blocks by solving cryptographic puzzles and the one who gets there first wins some newly mined BTC and transaction fees. Essentially the majority (50% plus) need to be working on the “correct” chain in order for it to be considered true.

This consensus requires a lot of energy to work, and fees are also considered very high, I see Seele having no problems surpassing this. In my view the clustering of mining centres for Ethereum and Bitcoin have made them less decentralised also which is another reason why I believe Seele will be a lot better than this.

Proof of Stake (POS)

With this consensus Investors stake coins in order to add blocks to the network. The chance of being picked to create blocks depends on the amount you stake (essentially your share of the coins). Unlike POW coins are not mined, the holders of coins already minted/mined decide on which blocks are valid. A block is created then the network nodes verify it and sign it. If a fork was to ever occur, holders spend their tokens voting on which one to back, as described by Witherspoon, (2018), those who vote wrong could lose their stakes also.

Decred and Peercoin are two projects which come to mind when we talk about this consensus, Jagati, (2018) suggests Ethereum will eventually change to this with their Casper protocol. However much of this has been talked of by Ethereum in a bid to tackle higher fees and scalability issues, I still believe they are a long way off.

I do like that this is more decentralised that POW and more energy efficient, however this does not have the best scalability and security properties. When I say security, I mean the fact that the “nothing at stake” problem which is widely considered to be POS downfall is a red flag.; this allows validators to essentially vote/stake on both sides of forks because there is very little computational power required. I believe Seele’s neural consensus algorithm has simple rendered this a thing of the past.

Delegated proof-of-stake (DPoS).

As described by Witherspoon, (2018), Daniel Larimer envisaged this, adding more security. EOS and Steemit are popular projects using this consensus. Essentially this works not by holders voting on validity of blocks but voting for delegates to do it for them. I believe there are 21-100 of these delegates in a DPOS system and are shuffled around when delivering blocks.

It is a great consensus because if delegates lag they will be voted out and replaced. This consensus has driven blockchain adoption further, however I feel this consensus is too centralised (small number of decision makers). This is scalable and energy-efficient, but not that secure. Seele offer a far greater consensus mechanism to this.

Byzantine Fault Tolerance (BFT)

This is another exciting consensus in my opinion, Stellar, Ripple, and Hyperledger are the most popular projects using this. It is extremely low cost, provides high throughput and is scalable. Essentially it involves “generals” and their “armies” deciding in unison on consensus of which there are two main types:

Practical Byzantine Fault Tolerance (PBFT)- I am not a fan of this method.
Federated Byzantine Agreement (FBA)- used by stellar and Ripple. With stellar anyone can become a validator but with ripple the generals are assigned by the ripple foundation. I am a huge admirer of this consensus algorithm due to its low cost and scalability and indeed its throughput.

With this being said I still believe Seele’s neural consensus algorithm is far more superior to this consensus also.

Other Consensus protocols which Cannot compete with Seele

Proof of Activity: This is a newer consensus which combines both POW & POS, essentially a hybrid approach. Mining begins as normal with POW but these blocks don’t contain transactions. POS then kicks in and chooses a random group of validators to sign the block; again, with a higher stake comes a higher chance of being picked to validate. At the end of the process when the block is created the fees are split between miners and validators. It is not that common, and it still requires a lot of resources for the mining element and is open to validator double signing, I feel Neural consensus is far greater than this.
Proof of authority: with this consensus transactions are validated by approved accounts. This is typically done with software and done automatically. Essentially it is a consensus which relies on reputation to become a validator. This allows very higher throughput and scalability and is a very exciting consensus' however I believe it is too centralised to drive mainstream adoption of blockchain. It is not as refined as neural consensus.
Proof of burn: This essentially involves sending coins to “Burn” for allegiance to the blockchain, this ensures your right to mine. The more you burn the more mining rights you get. Stakes eventually decrease over time so more burning is required which means continuous investment. This is not as widely used but has its merits, however neural consensus is in a league of its own when compared to this.
Proof of capacity: Most consensus require you to pay to play. However here you pay with your hard drive space – the more you have the more likely hood you have of being chosen to mine the next block. The more plots in your hard drive the more chance you have – So investing in hard drives is the key here. This is interesting but not widely used.
Proof of elapsed time: Intel came up with this according to Castor, (2017), similar to POW but requiring less electricity. Instead of participants solving the cryptographic puzzle the algorithm here uses a trusted execution environment (TEE), this ensure blocks are created via random lottery without the hard work. TEE gives wait time and is highly scalable. It is a great consensus as it is tackling the energy costs required to use POW, however Seele offer a greater concept.

Seele’s Neural consensus algorithm

E-Differential Agreement (EDA)

What makes Seele unique (4.0) is that they take from the many advantages of these current consensuses spoken of above and propose a new “ε-differential agreement (EDA). This agreement is essentially a mathematical process which will base their network on a single consistent value (i.e. something for the network to agree on). This is ground-breaking as EDA means every node would need to come to agreement on this value. According to SeeleTech, (2018) Seele will allow users to alter parameters to suit the use case:

The number of nodes that are sampled each round.
The target convergence interval, once all node values fall within the EDA the network is in consensus with itself, according to SeeleTech, (2018).
The number of rounds of voting.

Source: SeeleTech (2018).

What this all means is that once a transaction occurs it is distributed throughout the entire networks Nodes, and each of these nodes generates a pool of unconfirmed transactions. At this stage it is disorganised. Then a sample is taken from other nodes to record values which creates an aggregated value, according to SeeleTech, (2018). This creates the value statistic. Convergence then occurs (multiple sampling rounds), and once the difference between all nodes is less than the EDA then consensus occurs.

How neural consensus works

According to Seele, (2018) In their network the majority of nodes will not have to vote to reach consensus, instead they automate this process on chain via their algorithm. The more nodes there are online means it runs faster, interestingly their network will still work when there are nodes offline, it would just require more votes to reach consensus. I believe the limit for offline nodes is 33% which is still a lot and should provide a nice stable network.

Seele will do all of this whilst being compatible between different protocols such as POW and POS as mentioned above. It will be more secure also as random nodes are selected to execute these calculations. This image below shows their consensus model.

Source: Seele, (2018)

Advantages of Neural Consensus

According to Seele, (2018) The many advantages of Seele’s consensus model over the others are:

Their consensus process comprises of discrete voting rather than continuous voting making it more secure.
They will allow for their efficiency parameters to be adjusted for different environmental factors.
It will be a lot more energy efficient than the consensus listed above.
It will have very low transmission overheads thus making it cheaper.
Seele will allow for adjustable parameters to cater for different uses.
Finally, and most notable is that this is compatible with a variety of existing network structures.

What sets neural consensus apart

From what I can understand from their whitepaper this consensus allows for linear scalability (As node size increases so does performance). Essentially as the node size increases this allows for faster convergence and faster performance on the network. In their tests they tell us that when they applied a 100K node network environment their TPS reached 100K which is just mind blowing considering most of the consensus listed above cannot achieve even near this number. They also mentioned that their transaction confirmation delay decreased to several seconds.

This algorithm will handle all the data in an extremely efficient large scale decentralised environment. Their fundamental protocol utilises value internet alongside their heterogeneous forest network which connects not only their own chains but will allow them to interact with every other chain (BTC, ETH & 3.0 chains). What this all boils down to is a neural consensus algorithm in which when there is higher participation then the consensus convergence occurs faster.

As with many consensuses types out their validation occurs via nodes on the network and as a result of the high decentralisation it makes it a lot more secure. Also, this amazing team lead by Dr. Maolin Zheng (CEO) has ensured this innovative blockchain is far more efficient and fairer than its competition. What I love about Seele is that they can co-exist with various consensus algorithms and utilise benefits from them all. This is a complete ecosystem of technology which will allow for greater value of information transfer, thus enabling a more efficient society.

Conclusion

We all know the problems that blockchains face today- I.e. trying to maximise scalability whilst improving decentralisation and security. I believe Seele have found a solution to this with their neural consensus protocol. I have touched on this time and time again, scalability is what most blockchains are competing against each other for today, but many seem to be sacrificing decentralisation in the process. As a network becomes more centralised it can scale better; however, it is the projects that can scale whilst also retaining a high degree of decentralisation and security which have my vote of confidence, Seele is such a project.

Seele’s consensus algorithm tackles these problems directly and offers much more value than their competition. They achieve this brilliance via their ingenious neural consensus algorithm and E-Differential process. They are far superior to the current consensus protocols used today. One just needs to compare them like I have done so here to see why.

Seele is currently doing very well on the exchanges and I believe it speaks volume to the calibre of project we have here. I would urge everybody to do some research on this project and see for yourself.

References

Castor, A. (2018). A (Short) Guide to Blockchain Consensus Protocols. [Blog] CoinDesk. Available at: https://www.coindesk.com/short-guide-blockchain-consensus-protocols/ [Accessed 18 Jun. 2018].
Jagati, S. (2018). Ethereum’s Proof of Stake Protocol Under Review. [Blog] Crypto Slate. Available at: https://cryptoslate.com/ethereums-proof-of-stake-protocol-in-review/ [Accessed 18 Jun. 2018].
Keeping Stock. (2018). Explaining blockchain — how proof of work enables thrustless consensus. [online] Available at: https://keepingstock.net/explaining-blockchain-how-proof-of-work-enables-trustless-consensus-2abed27f0845 [Accessed 18 Jun. 2018].
Medium. (2018). Seele Blockchain 4.0? EOS Killer? – Omnis Crypto – Medium. [online] Available at: https://medium.com/@omniscrypto/seele-blockchain-4-0-eos-killer-7a69d34713bf [Accessed 26 Jun. 2018].
Seele.pro. (2018). Seele. [online] Available at: https://seele.pro [Accessed 26 Jun. 2018].
SeeleTech (2018). Blockchain’s New Brain: An Introduction to Seele’s Neural Consensus Algorithm. [online] Medium. Available at: https://medium.com/seeletech/blockchains-new-brain-an-introduction-to-seele-s-neural-consensus-algorithm-9963749f03b5 [Accessed 26 Jun. 2018].
Witherspoon, Z. (2018). A Hitchhiker’s Guide to Consensus Algorithms – Hacker Noon. [online] Hacker Noon. Available at: https://hackernoon.com/a-hitchhikers-guide-to-consensus-algorithms-d81aae3eb0e3?gi=669e3bb01191 [Accessed 26 Jun. 2018].

Why Seele’s Neural Consensus Algorithm is superior to its competition