Those familiar with market failures and blockchain may find the Saito Whitepaper a more succinct guide.
Saito consensus primarily addresses two well studied market failures, The Free Rider Problem , and The Tragedy of The Commons - both well-founded and recognized game theoretic dilemmas; the only assumption being that self-interested actors behave rationally.
Market failures are situations in which the rational incentives of individuals leads to less prosperous outcomes for everyone. These sub-optimal outcomes are pervasive due to the extreme sensitivity of naive systems to 'defection;' few dishonest players can exploit the honesty of the group and turn the 'honest' or 'nice' strategy into a losing one.
This dynamic is commonly referred to as a Prisoner's Dilemma; the Wikipedia link lists some notable real-life examples. The standard reaction to the dilemma in the context of market failures is to privatize access to the good in order to protect against exploitation.
Saito exists to solve market failures in Proof of Work and Proof of Stake, without privatization of the blockchain. Addressing these market failures is crucial for scalable, secure, sustainable and most importantly, open, blockchain.
Saito is the first and only layer-1 blockchain to address these market failures on the consensus level, thereby avoiding the tough decision between remaining open or scaling network bandwidth. Saito consensus aligns the incentives for scale and openness as a single fee-sharing mechanism, and avoids collapse via bloat through a market-based rent system.
The cryptographic schemes in Saito which are responsible for addressing the market failures in Proof of Work and Proof of Stake are relatively simple, certainly less complicated than many modern Proof of Stake consensus protocols; so it is natural to ask: why is Saito the first and only blockchain to address these market failures?
The answer is that merely recognizing the true nature of the problem is difficult. Seeing that issues around scaling, openness, and security can be distilled down to more fundamental problems and understanding how exactly cryptography and incentives can be employed to address those problems is where the sophistication lies. Once the problems are clearly identified, the mechanisms required to address emerge somewhat naturally.
Thus in attempting to understand Saito, many study the cryptographic schemes and understand them on a technical level, but still cannot truly explain what motivates the incentives therein. Most blockchain issues are framed as technical problems, but are in fact incentive-level, game-theoretic, economic problems - therefore many people study Saito with the wrong frame-of-mind.
The problem with blockchain scaling is not at the network technology layer: at the time of writing, data centers around the world are im- plementing 400 Gbps network switches while 100 Gbps connections are becoming standard even in lower-tier colocation facilities...
...What limits network growth is the challenge of paying for the network. In the past, non-economists have waved away this limitation, claiming that as long as someone is earning money from the network they will pay all costs necessary to support it. But this is not true...
-Saito Whitepaper
This page hopes to explain the incentive-level issues which lead to market failures in Proof of Work, Proof of Stake and other non-Saito blockchains. The technical solutions may be easy to understand, but the economic motivations informing those solutions are key to a holistic understanding of Saito and blockchain generally.
Free Rider Problems occur between the funding of a non-excludable good (one that everyone can enjoy) and the benefit of that good. Consider a group dedicated to removing litter at a public nature-park: everyone collectively enjoys the benefit of a cleaner park, but only those donating time or money to the group have to pay for it. Everyone else is allowed to 'free-ride' off the expense of the maintainers.
Markets need to control the distribution of benefits to induce people to pay for them. If you're producing something that everyone can enjoy, you'll have difficulty convincing someone to pay you to cover the cost.
-Saito FAQs
The classic solution to this problem is to privatize the once public good - to exclude access to its benefits in order to solicit appropriate payment. But this solution is inadequate for goods which can't be restricted from public use, or which have open access as a founding principle; in the case of blockchain, the routing of fee-paying transaction towards the block producers who hoards the total fee-reward is a Free Rider Problem.
One of the key innovations in Bitcoin was that the act of securing the longest chain by mining atop it was not to be restricted to any authorized party, but to be a responsibility available to and rewarding to anyone; this is crucial for the incorruptibility of the network.
Likewise, any user wishing to send transactions need only to propagate that data to a node, or set of nodes, capable of mining it into a block - at least that was the vision. In theory and in practice, the collection of fee-paying transactions in PoW and PoS results in a Free Rider Problem, because sharing transaction data means giving up the ability to earn the fee.
On Bitcoin and Red Balloons famously and correctly identifies this problem within blockchains which existed at the time. However, it incorrectly proports that "there are no reward schemes in which information propagation and no self-cloning is a dominant strategy - " (skipping ahead) this conclusion is proven incorrect using Saito consensus in A Simple Proof of Sybil Proof.
Miners who wish to make money off of transaction fees must possess those transactions within their mempools before they begin mining. Since it can't be known who will produce the block, transactions must be sent to an arbitrary and large set of miners if they are likely to be included as quickly as possible.
Because nodes or other services which collect and transmit transactions do not know who the next miner will be, they must send their transactions and the right to earn their fees to all miners; the nodes who actually route fees to the block producers are unable to earn those fees. The block producers (miners, stakers) are Free Riders atop transaction-routing nodes, because transaction relayers have no choice but to give up these transactions if they are to remain useful to their users.
In order for transaction collection to happen at scale on a PoW or PoS blockchain, the infrastructure which collects transactions must privatize that service in order to ensure they receive payment. This payment is either an additional cost users must pay, or it detracts from the 'security budget;' i.e. the fees which grant the blockchain economic security against consensus attacks.
This inadvertent privatization can turn large portions of a blockchain network into a permissioned system, but the other option is to remain a non-scalable chain. The debates around "scale versus decentralization" can be more precisely framed as tradeoffs around an unaddressed Free Rider Problem.
Different blockchains cope with the Free Rider Problem in different ways. Bitcoin very famously chooses to forgo scaling in order to ensure unpaid, low-powered nodes can and will route transactions to miners. Because the miners are expected to free-ride off of these nodes for the network to function, the chain has chosen to remain small with the hopes that volunteers will continue to perform this task without compensation.
Ethereum is a great example of the opposite tradeoff. In pushing the Free-Rider Problem towards its natural conclusion, crucial network infrastructure tends towards private control and centralization.
The flagship example is the API node service Infura which is majorly responsible for serving applications both in service of developers and users. The massive userbase of the Metamask wallet is primarily routed through and dependent on Infura, not to mention the prolific share of 'decentralized' applications reliant on it. Note that Infura and other node providers mainly run their infrastructure atop even more centralized operations like AWS.
Infura and similar node providers serve more than just Ethereum. Dozens of chains with the ambition to scale their blockchain like or past Ethereum quickly outpace what independent full nodes are able to provide and fall-back to private solutions like Infura. This dynamic concentrates power and profit to private companies which can censor, and spy on users, and which take security from the underlying blockchain to instead finance closed revenue streams.
Private operations running blockchain nodes isn't the issue, but their lack of accountability towards the blockchains they serve is. Whereas mining and staking have built-in punishment for nodes who misbehave, censor, or perform sub-optimally, the typical dynamics of monopoly are the rules that apply to node services whose work is so disconnected from consensus.
Recall the dilemma transaction-serving nodes face: in order to serve their users well (get transactions included in next block) they must distribute transactions to as many potential block producers as possible, but in order to profit from those fees, they must restrict access to that transaction data; they must privatize. Since block producers are necessary to fulfill the service, most node providers choose to privatize access to user-facing service instead.
In Saito there is no such dilemma. Nodes which share transactions do not risk the block producer who picks them up 'stealing' it from them - instead, the nodes with earlier access to the transactions become entitled to larger shares of the fee and compete with other early nodes to route the version of the transaction with their claim on it first. The incentives become aligned: nodes are both rewarded for collecting and sharing transactions.
The mechanism which solves the Free Rider Problem is Routing Work. A transaction carries a chain of digital signatures demarking how much fee each relay is due and how quickly each relay can produce the next block.
Because public-facing relay nodes are no longer doing free work for block producers, the infrastructure of the network which connects users to the blockchain is not required to restrict access in order to profit, thus the network layer can be funded without defunding security. This means scale and security are not mutually exclusive, they are in fact the same metric in Saito - therefore, Saito proves the blockchain trilemma does not exist.
What users value and what nodes value in Saito is the same. Users want fast transaction confirmation (and security), and relay nodes secure the network and earn fees by getting their version of a transaction into a block before a competing routing node does. Because consensus nodes and users have the same incentive, fees pay for what users actually value and optimizing rewards means optimizing user experience.
The practical benefits which stem from this fundamental shift in incentives are profound:
The Tragedy of the Commons (TOTC) occurs when a valuable, publicly-accessible resource (non-excludable good), is quickly depleted as competing or self-interested actors race to consume it. The classic example is a public field which farmers compete to freely feed their livestock until it is barren.
The comic above shows a more abstract version, where the quality of public air is degraded in order to increase production; the cost of a degradation in air quality is paid by everyone equally, rather than solely the polluters. When the air quality is so poor that people are too sick to work, the resource may be considered fully consumed.
Despite the fact that the depletion results in a situation where everyone collectively is worse off, each individual during the process has a rational incentive to continue consuming the resource, and to do so with greater effort than others. As with Free Riding, the classical solution to the problem is to relegate control over that consumption to a private interest.
The Tragedy of The Commons in blockchain is more similar to the pollution example, since long-term storage burdens can be incurred on blockchain for a one-time price:
The tragedy-of-the-commons issue is created by the existence of the permanent ledger, which encourages nodes to accept payment today for work that can be offloaded to others tomorrow. This incentive leads to bloated blockchains and more subtly to transaction mis-pricing, as users can pay fees that do not reflect the true cost of their transaction to the overall network.
-Saito Whitepaper
There are a couple ways rational profit-seeking nodes will end up exploiting the storage commitments of more 'loyal' nodes. The first is a node which produces blocks, earns fees, but does not bother to validate or even hold historic transactions. For blockchains with large blocks, it may be costly for anyone to validate a block, and the job may end up pushed to just a few nodes who copy each other's work.
The second way long-term nodes can be exploited by short term interests is from nodes who come online during a surge of high demand in order to earn the fees of that surge. The massive influx of new transactions will pay the nodes as soon as they are added, and when the surge subsides, those nodes will take their profits and leave the data for everyone else to hold and maintain.
The more subtle version of both of these issues is blockchain bloat: the network slowly accumulates data past what it can support without expenditure; when nodes begin spending more money on storage, it leads to new transactions paying the debt of storage incurred by the old ones.
Even if hardware storage prices continue to decrease, without a sound economic model to price blockchain storage, fees will either be higher than necessary or too low to support existing infrastructure. A true market solution will render lower user fees as storage costs decrease - not simply maintain some set level of efficiency.
Arweave is one project based entirely around storage, and which uses an economic model which attempts to predict the longterm price dynamics of digital storage. Not only will this lead to mismatches between the real cost of storage and the model's assumed cost, but if the model needs to ever be adjusted then it will again require developers making decisions about incentives and how node operators are paid.
Arweave is not alone in relying on developers to attempt to predict storage costs for the chain. Most big-block chains make even less rigurous assumptions: namely, that since storage costs decrease over time that accurately pricing it isn't an issue. This is common in big-block forks of Bitcoin.
But even more ambitious chains at least do recognize that continually adding permanent data is not sustainable. The solution for them is to not use the blockchain - literally, their answer is to push transactions onto less secure and more centralized layers 2, 3 and above. Because it is assumed there are no solutions to deal with this problem on the layer 1, most smart contract chains believe that most data shouldn't enjoy the benefits of the layer 1.
And again, holding the uber conservative side of the spectrum is Bitcoin. Bitcoin hopes that by keeping blocks very small, any economic problems which occur out of a model which leads to a Tragedy of The Commons Problem will be insignificant. It's as if Bitcoin is saying - without a solution to this problem, we will not scale. Saito identifies the problem and poses a solution.
In Saito, not only is the price of storage accurately adjusted automatically and in real-time, but transactions by default pay for the minimum amount of storage required, rather than paying for the maximum (forever). This is because transactions in Saito are allowed to expire after their epoch, but if they have a live balance, they must be rebroadcast into a fresh block such that the UTXO set only relies on blocks from the past, while also paying a fee.
Hence the name: Automatic Transcation Rebroadcasting (ATR). Once any unspent transaction is N blocks old, it must be rebroadcast and will automatically pay a fee from its balance when it is. If the transaction does not have sufficient funds to pay the fee, then it is removed from storage and nodes are no longer responsible for it. The fee paid is a positive multiple of the average fee paid over the transactions lifespan.
This benefits users and full nodes equally. Because a node cannot produce a valid block without including all transactions which must be rebroadcast, a node cannot produce blocks until it truly becomes a full node and is holding the full UTXO set and the blocks which validate that set. This makes Data-Availability a default incentive for Saito.
Because full nodes are now paid for holding transactions, they do not care if data sticks around forever, because it will also pay them forever. On the flipside, users are no longer required to pay upfront fees which assume that data will be around forver. Because transactions can be removed, user fees only need to pay to add data - if it sticks around for rebroadcasting, it will pay for storage later.
This is a massive boon for use of the blockchain as a messaging layer. Posting data on any chain which does not have ATR means paying upfront for forever-storage costs. On Saito, transactions only need to pay for their first epoch when they are added, which means greatly reduced transaction fees and a greater ability to use the blockchain to transmit messages universally around the globe - making Saito an ideal PKI network.
By cmglee, http://github.com/emojione/emojione/graphs/contributors, http://github.com/twitter/twemoji#committers-and-contributors - This vector image includes elements from this file:, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=140014761