2017 was the year we saw different projects pioneering the blockchain technologies. Not confined in the Bitcoin blockchain, various teams embarked on Ethereum development, on NEO, Qtum and many others. In the meantime, variations of consensus were introduced to adapt to specific applications, adjusting and balancing among fairness, security, participatory, scalability. In this article, l will have a brief overview of several types of consensus and how they differ from each other. Most importantly, to explain Qlink’s new initiative — Shannon Consensus.
To start with, let's have a look at how consensus reshaped the traditional finance industry and beyond. Bitcoin disrupts the value transfer business model, created a mechanism of all network participation in ledger keeping. Most importantly, it provided the option of not using the banking system or traditional money transfer companies who usually charge high transaction fees. Take a step further, we can imagine various computation methods for non-middleman coordination, to name a few, logistics, arbitration, clearance. This is fascinating when thinking about how to balance the resources consumed to maintain the public ledger with the P2P security, fairness that decentralization provides.
POW
In early 2015, The Economics published an article complimented the blockchain technology “The Trust Machine”. The article described Bitcoin's blockchain consensus mechanism, called Proof-Of-Work (PoW), that solves an important problem in distributed computing and that is the consensus problem. One of its famous instances being the Byzantine General's Problem. Revolutionary Proof-of-Work allows a trustless network of nodes to come to an agreement, tackling this way the double-spending problem (link to https://en.wikipedia.org/wiki/Double-spending). Handling the double-spending issue was the first fundamental question for any digital currency to work.
PoW permits a single participant to announce their result (block of transactions), which all other participants have to verify, under the condition that a complex mathematical puzzle was solved. These complex problems that the participants have to solve are based on a hashing function. Incorrect or malicious results are prevented by the fact that all other nodes participate in the verification and by the limits imposed on the hash function result.
PoW, though trustless, permissionless and ideally secure, has been criticized for its native problems, including a lack of storage, extremely electricity consuming, long confirmation time, not to mention the concerns that if one mining pool or several ones work with each other own over 51% controls of the computing power, the whole mechanism will be compromised. The advantage of Bitcoin still stands, as the most distributed value transfer system, however, the question remains that when the hype goes away, and price stables, when and what is the best application scenarios for Bitcoin usage.
POS
Proof-of-Stake (PoS) algorithms solved the computing power consumption problem by replacing the hash function calculation with a simple digital signature which proves ownership of the stake. Instead of an individual to conduct the value calculation in order to be added to a block, the network itself runs a lottery to nominate.
Consequently, the ones who own more tokens will accumulate their possessions step by step, in another word, become “richer”. The concentration of voting power increases as PoS’s major drawback. The richer ones, in this sense, are comparatively easier to conduct a “fork” attack. From the appearance of this consensus, assuming participants' egotistic, it seems that participating in PoS inevitably becomes a game for power accumulation, gaining weight in a decision making.
IMPROVEMENTS
Delegated PoS (dPoS) requires a validator has comparatively larger stakes and the mass participants who hold fewer tokens can delegate their assets to the wealthy ones. This is a solution towards efficiency improvement while sacrificing a certain degree of trust by adopting the indirect delegation mechanism.
Byzantine Fault Tolerance (BFT) algorithms remove the validation process and apply a coordination system, means BFT is looking for a solution which the most participants agree on. Similar to dPoS, BFT also increased the block generating speed to a greater degree. However, the participatory appears a problem as participants tend to ignore or blacklist potentially malicious ones.
Directed Acyclic Graph or DAG algorithms was born to solve the scalability problem. The algorithm uses a Directed Acyclic Graph (DAG) data structure instead of the single-threaded blockchain data structure that the other algorithms use.
The consensus mechanism first brought up by Satoshi Nakamoto successfully inspired more creative ideas. The starting point is to seek agreement while minimizing the chances of individuals sabotaging the collective achievements when putting their own interest in front. It’s an experiment of using computer science and mathematics to balance group interest and individual ones.
New proof-of-everything is pumping up every now and then, competing to be the “perfect” one. Some fixed the transaction speed part, some the scalability part, others focused on the wealth concentration part. Some sacrificed the egalitarian rule, others sacrificed participatory rule or inclusive rule. New mechanisms equally intriguing are proof-of-burn, proof-of-capacity, proof-of-eclaped time and more. Eventually, what we are looking at shall be the applications as it is highly unlikely to have an omnipotent consensus rule that caters every scenario.
To adapt to the communication service industry, Qlink introduced Shannon Consensus:
The Shannon consensus is dedicated for network transmission. For network services and applications, on average 80% of the jobs for network transmissions will be done by 20% of the nodes. This algorithm improves the Proof of Work consensus by separating the working nodes from the ledger nodes, and encourages more inactive nodes to participate.
The “rich” working nodes will only be used for transmissions, while the nodes with less work, called “middle-class” nodes, will be assigned to bookkeeping. Rich nodes have more tokens and more stake in consensus than the middle class nodes. The rich nodes will be rewarded by handling the transmission tasks, however, they are asked to share the rewards with the “middle class” who manages the ledger. This enables a more reasonable profit sharing plan and creates a fairer ecosystem.
The consensus uses unique proof mechanisms, called Proof of Transmission, Proof of Spacetime and Proof of Retrievability so that the actual work will be based on the effective workload of transmission. Furthermore, the process is the cooperation between multiple nodes rather than a single node, and cooperative nodes acting together will make the network more secure and trustable.
This consensus still needs validation and more projects to adopt and trial. However, the concept is rather interesting knowing that a fairer economy building is first time come up as the core mission of a consensus.