Pi Futures: Understanding Stellar Consensus Protocol, and why it takes so much time to ground.
The Stellar Consensus Protocol (SCP) was first introduced and architected by David Mazières, a professor of Computer Science at Stanford University, in 2015. SCP has been implemented in Stellar project, and now, it also plays a core role in Pi Network. The protocol uses an algorithm called Federated Byzantine Agreement (FBA), an alternative to the Byzantine Fault Tolerance (BFT) algorithm. Its biggest improvement is to allow anyone to join the system and become a message validator, instead of allowing only a fixed group of validators like the original version. This makes FBA more suitable for a decentralized system.
The best way to understand the Stellar Consensus Protocol or FBA is to learn through its pros and cons.
The Stellar Consensus Protocol helps to establish an open membership network, and enables the messages between the network nodes to be disseminated in an resource-friendly and efficient manner. It means SCP network is decentralized and runs fast. But how does it achieve that?
In a BFT environment, there must be a recommended message validator list defined by a central authority. Anyone can spin up a validator, but you can only participate in consensus if the authority adds you to the list. This requirement for a recommended validator list means that BFT is a centralized enclosed membership system. In proof of work (POW), like Bitcoin, anyone with a mining machine can participate in consensus; and miners can join and leave the network without impacting consensus. But it takes much time for POW network to reach consensus. In SCP or FBA, there is no recommended validator list chosen by a central authority, rather each validator decides which other validators they trust. And their list of trusted validators is called the quorum slice, which is known as the security circle in Pi Network. The quorum slices of each validator overlap to form quorum or network-wide consensus on a transaction. Anyone can spin up a validator and participate in consensus, if any other participating validator adds you their quorum slice. So to summarize this, quorum slices allow for open membership and therefore decentralization, as well as better the running efficiency of the network.
However, when building an asynchronous distributed system, it is necessary to find a balance between openness, operational efficiency, and safety. In other words, you cannot create a network that is completely decentralized, while it is energy-saving, runs at high speed, and maintains a high degree of data consistency.
The safety feature means the mechanism that maintains data consistency in the network. The open network must be able to tolerate nodes that lie, or send incorrect messages, or confirm ledgers with false transactions. In SCP, if the network cannot agree on the ledger and faces the event of an accidental fork, it halts progress of the network until consensus can be reached. In this case, most likely some authority has to go in and manually solve the problem. This will result in that the network is not autonomous, or is unfeasible and unsustainable.
Blockchain technology came out with Bitcoin. In a completely decentralized network environment, the blockchain ensures that there is one and only one ledger among nodes. Although the theory of Stellar Consensus Protocol has become more and more perfect, it is not a blockchain after all, but a protocol and algorithms used in distributed systems. How to apply it to social scenes and mobile devices is a major challenge for Pi Network in the future.
Pi Futures engineers and technology enthusiasts are so ready. When Pi Network releases its source code, we will invest in research and development in Pi Network’s ecology.
Pi Futures is now publicly traded on Uniswap v3. The current price is $1.34. Trading Link: https://info.uniswap.org/#/pools/0x4351825d4b6b92264c76ffa10b4981074fa01326
Disclaimer: Pi Futures (symbol: Pi) is ERC-20 token, and is neither the same nor the alternative of the cryptocurrency of Pi Network.