Plasma protocol for newbies

In September, Joseph Poon (founder of Lightning Network) and Vitaly Buterin (founder of the Ethereum cryptocurrency) introduced the first draft description of the Plasma protocol (https://www.plasma.io), designed to significantly increase the transaction volume of blockchain networks to the tune of billions of operations per second, according to the authors themselves.
Disclaimer: the description given below is an attempt to explain the principles proposed by the authors in simple terms. It does not, in any capacity, reflect the complex series of actions and motivations of players, which must be analyzed by game theory experts. The description is also currently a work in progressand is in many ways incomplete.
The approach of the authors is largely similar to the Lightning Network idea — the execution of transactions off-blockchain by freezing funds in the base network and using them for out-of-network settlements. The basic characteristic of out-of-network settlements is the lack of a centralized settlement service, and the use of a pre-determined interactive protocol directly between participants of the settlement (the sender and the recipient, whom we will also call Alice and Bob). Alice and Bob themselves have minimal trust in each other and are, therefore, involved in a certain kind of “game” — if one of them catches the other behaving inappropriately (breaking the pre-determined rules of the transaction), then the other participant of the “game” may submit proof of cheating to the parent network, causing the “cheater” to be fined.
In the Plasma protocol, this idea is developed and generalized. An entire new blockchain is created off the root blockchain! The consensus rules in the new blockchain as well as the behavior of participants and the functions to be performed are described by a smart contract in the root network. The state of a potentially large number of participants in the daughter network is encoded as a compact submission with the help of a Merkle Tree and recorded in the parent blockchain, allowing external observers to follow the state of the daughter blockchain. In order to participate in the daughter network operations, participants must first send funds to the daughter network smart contract.
First, let’s suppose that a Proof of Authority (PoA) consensus protocol is active in the daughter network — only a certain set of nodes may confirm transactions. The operator — the creator of the daughter network smart contract — is responsible for updating data and pays significant funds as collateral into this contract. These funds constitute an incentive for the operator to obey the rules. Since all network participants (both, those participating in the parent blockchain, and potentially external ones) can observe transactions in the daughter blockchain, if they find a mistake in the settlements and submit what the authors of the article call “fraud proof”, then, according to the rules of the smart contract in the parent network, the operator will be fined a significant amount from the funds it has initially deposited, and users of the daughter network will be able to return their funds to the parent network.
This protocol is wholly interactive, and Alice and Bob are ideally supposed to follow ongoing transactions within the daughter network and make sure that they are correctly added to the daughter chain, and that the state of the daughter chain is correctly transmitted to the parent blockchain. The time window of opportunity for Alice or Bob to catch the blockchain operator cheating is final, since, quote: “whereby if a tree falls in the forest and nobody listens to it, it presumes that it doesn’t matter whether it makes a sound or not”.
The proposed architecture allows us to circumvent the basic limitation of the root blockchain — a fairly high transaction cost and the cost of information storage in the chain (at least in the ongoing Proof of Work (PoW) protocol), since the state of the daughter chain is compactly submitted to the parent, and the calculation of internal states takes place within the daughter chain. The parent blockchain nevertheless functions as a kind of supreme arbiter over the daughter chain, allowing daughter blockchain participants to bring their funds back out in the event of inappropriate behavior.
The daughter network can also have other consensus mechanisms, the authors also propose a variation of the Proof of Stake (PoS) protocol, although they recommend treating their variation with a certain amount of caution. In addition, the daughter network does not need to operate with the base currency of the root network, ethereum (ETH) — it can use any ERC20 compatible tokens. Since the value of these tokens, in relation to other units, will also be determined based on the correctness of participant behavior in the daughter network, it constitutes an additional incentive for all participants to behave correctly.
You might ask — how does this increase the speed to a billion transactions? The daughter networks are, after all, subject to many known limitations, even if they might potentially lower the transaction cost and processing speed. To solve this, the authors propose going deeper and creating another daughter within the daughter network!
This creates a hierarchy of blockchains, and every higher blockchain acts as an arbiter and looks after its daughters. Of course, the authors explore various marginal situations when the “daughter” is well-behaved, but its direct parent blockchain breaks the rules.
Such a hierarchy of blockchains largely echoes the idea of the map-reduce architecture, which is proposed to be used for the realization of arbitrary settlements with the help of a blockchain tree. The article offers examples of the realization of a Reddit analogue, decentralized exchange, electronic mail and a Content Delivery Network (CDN). Nonetheless, this approach to architecture building does not allow for the realization of all arbitrary business processes, which the authors also recognize.
This document constitutes a draft and will be considerably supplemented, with the description expanded. Nonetheless, it is already clear that the proposed protocol is still very far from being implemented, although in some time we may definitely see some companies realizing similar functionalities for commercial goals. The architecture description may also contain hints that the Ethereum base network will not switch to the PoS consensus protocol, but that a daughter network with the protocol will be created instead.
More information is available at: https://bankex.com/
Telegram: https://t.me/bankex
Facebook: https://www.facebook.com/BankExchange/
GitHub: https://github.com/BankEx