Uinspire Blockchain Course {Lesson 5: The Future of Distributed Ledger Technology}
In Lesson 4 we focused on three innovative applications that are powered by smart contracts: ICOs, DApps and DAOs. These applications are at the forefront of most of the current discussions regarding blockchain technology.
In this Lesson, we aim to provide an insight into recent and future developments within the world of DLT that may have a significant impact on this space. Building on the previous Lesson, we start by examining how DApps may collectively give rise to the distributed economy. We will then cover a range of exciting topics, including novel forms of fundraising, alternative consensus mechanisms and new DLT structures.
The Distributed Economy
The goal of most DApps (and DAOs) is to coordinate transactional activities between a large set of unknown actors and to reduce transactional frictions by removing central intermediaries. In order to achieve this goal, DApps are able to communicate with one another by passing on messages between their underlying smart contracts. The diagram below depicts these communications taking place between DApps on Ethereum, though much research is also being conducted in relation to cross-chain communications.
In the long run, the increased interconnectivity of DApps will help to empower the distributed economy. This economic model will be characterised by the absence of central intermediaries and the automatability of communications and tasks associated with the negotiation and performance of virtual agreements. DApps will be able to exchange payments, information and products with minimal human intervention. This can be contrasted with the sharing economy, in which corporations that match service providers to consumers charge significant fees to satisfy their investors’ desire for financial returns.
Let’s take the example of insurance. At the macroeconomic level, Ethereum provides the bottom-layer protocol which the DApps that run on top of the network must follow. At the microeconomic level, the figure above shows how a collection of DApps can offer insurance-related services to end users. In particular, users could interact directly with the Insurance DApp to obtain a particular policy to protect themselves against an identified risk. This Insurance DApp would then rely on uPort to perform the necessary Know-Your-Customer (KYC) checks before onboarding the user. The Insurance DApp could then use Golem to obtain crowdsourced predictions about the probability of the relevant risk materialising and request external pricing data from other insurance companies using Oraclize. The predictions and external pricing data could then be passed on to Golem’s distributed computing network to calculate an appropriate price for the insurance policy. With the addition of Aragon, the Insurance DApp can integrate an off-the-shelf governance system to process insurance claims and deal with disputes.
By way of caution, we are closer to the scenario of The DAO than to the vision of the distributed economy outlined above. Though we are still in the early stages of smart contract development, many resources are being deployed to achieve progress in this field. We can expect the rate of progress to accelerate as DApps become fully functional and the learning curve associated with blockchain technology begins to flatten.
Alternative Forms of Token Distributions
Over the past year, ICOs have come under the scrutiny of regulators from all around the world. This has raised serious concerns about the legality of ICOs, as project founders may have failed to comply with the applicable securities legislations.
Despite the evolving regulatory landscape, the sums raised via ICOs have carried on increasing in 2018 (compared to 2017). Nevertheless, many industry participants have lost confidence in this fundraising method after facing countless scams, projects with no track record, opaque information and the lack of a minimum viable product (MVP). As a result of these diverse tensions, a number of alternative token distribution mechanisms for DLT projects have started to gain traction.
Airdrops
In an airdrop, tokens are distributed directly to users without them having to contribute funds in exchange. This is typically done by allowing users to register their Bitcoin or Ethereum addresses on the project’s website in order to receive a fixed amount of the project’s tokens. Alternatively, some projects require users to perform certain actions before they are able to receive the new tokens, such as registering on the project’s Telegram group.
Given that the token issuers do not raise any funds in this process, the risk for opportunism and scams is far lower with airdrops than with ICOs. Users who receive tokens via an airdrop still have an incentive for the project to succeed, as they can expect the price of the tokens to grow as the demand associated with the project increases. Others may be interested in the utility of the tokens and may plan to use these once the project becomes functional.
The SAFT
The Simple Agreement for Future Tokens (SAFT) was developed by CoinList, a token sale platform that whitelisted projects can use to raise funds. The purpose of the SAFT is to allow a project to raise funds without having to distribute tokens at this initial stage. Instead, the tokens will be distributed at a future point in time. This is usually once the platform becomes reasonably functional. At this point, it is believed that the investors will gain the ability to sell the tokens received without breaching the applicable securities legislation.
The SAFT draws inspiration from Y Combinator’s Simple Agreement for Future Equity (SAFE), which is frequently used by seed round investors. The SAFT itself is treated as a security under US law, which is why exemptions are relied on to avoid complying with the onerous regulations that apply to the sale of securities to the public. As such, CoinList restricts the use of the SAFT on their platform to accredited investors, which are exempt for these purposes. However, other projects such as WeFunder have used their own SAFT that allows a limited number of non-accredited investors to participate, relying instead on a different regulatory exemption.
The usefulness of the SAFT as a means of achieving regulatory compliance has been questioned by legal scholars. The regulators have not yet pronounced themselves on this topic, though the amount of capital that is being raised privately by using SAFT is on the rise.
Note: a similar concept, known as the SAFTE, was developed earlier by Jack du Rose from Colony.
Alternative Consensus Mechanisms
In the realm of DLT networks, concerns have been raised about the vast amounts of electricity that PoW systems consume. For instance, Bitcoin is said to currently consume as much electricity per year as Chile. Some debate the accuracy of these comparisons and argue that the banking system consumes far more electricity than Bitcoin. However, the banking system serves billions of people and allows a wide variety of financial services to be offered beyond the processing of payments.
In addition, we learned in Lesson 2 that PoW systems have proven unable to effectively combat centralisation. In Bitcoin, a small number of mining pools dominate the block production process. Ethereum introduced an algorithm that prevents those who purchase specialised mining hardware from gaining an advantage, though significant levels of concentration can still be observed. There is still some heated debate in this regard, with recent research suggesting that the networks may be less centralised than initially apparent.
In addition to these concerns, it appears that PoW is inherently hard to scale. We saw in Lesson 2 and Lesson 3 that the throughput of Bitcoin and Ethereum is capped (at approximately 7 TPS and 14 TPS respectively). As a result of these issues, much emphasis is currently being placed on finding a more scalable consensus mechanism.
NB: miners (and mining) are terms that are typically associated with PoW networks. The umbrella term that can be used across different consensus algorithms is block producers. The term block validators is sometimes used in this context too. However, this can be confusing as every node in the network is usually in charge of validating new blocks (as opposed to producing them).
Proof-of-Stake
Used by: Peercoin, Cardano, Ethereum (soon)
Proof-of-stake (PoS) is a consensus mechanism that has recently become popular amongst DLT networks. In a PoS system, nodes can stake (i.e. lock up temporarily) the network’s native currency in exchange for the chance to produce a new block of transactions. There are several variants of PoS algorithms, though a node with X% of the total staked tokens will generally have X% chance of producing the next block.
It is often argued that PoS systems are better at combatting decentralisation, as it is usually more expensive to collect X% of the network’s currency than it is to amass X% of the total hashrate. However, much depends on how the network’s currency is initially distributed. Critics have also observed that PoS systems will favour wealthy users, though the same can be said about PoW in practice.
Delegated Proof-of-Stake
Used by: Steemit, Bitshares, Lisk, EOS
Delegated Proof-of-Stake (DPoS) was invented by Dan Larimer and differs from PoS significantly. In DPoS systems, tokenholders vote to elect delegates that will serve as the network’s block producers. The number of block producers in a DPoS system is fixed and can range between 21 (as in EOS) and 101 (as in Lisk). This allows the delegates to organise themselves effectively and create designated time slots for each block producer to publish a new block. If a delegate continuously miss their slot or act maliciously, tokenholders will be able to vote for a new block producer to take that delegate’s position.
Some have criticised DPoS by arguing that the consensus mechanism is centralised in nature. However, centralisation will not necessarily be an issue in circumstances where those who have special administrative powers can be replaced easily if they misbehave or underperform. In other words, contestability (the ability to replace delegates or agents) can help to manage centralisation and prevent abuses from taking place, while still leaving room for scalability.
Federated Byzantine Agreement
Used by Ripple (permissioned) Stellar (unpermissioned)
In a Federated Byzantine Agreement (FBA) system, a minimum fraction of nodes (the quorum) is required in order to achieve overall consensus. However, nodes can choose to trust only a limited number of other nodes, thereby achieving consensus within the relevant “quorum slice”. For instance, Node A can choose to trust a new set of transactions only if Node B and Node C both agree on the new transactions. These quorum slices should have some level of overlap with others, allowing overall consensus to emerge.
Note: when quorum slices do not intersect, the system ends up with disjoint quorums. This is undesirable as it slows down and undermines the overall consensus.
New DLT structures
In Lesson 1, we learned that blockchain (in the narrow sense of the word) simply refers to a type of data structure. Other DLT structures have recently gained popularity to to deal with the scalability issue that is associated with blockchain (and the associated consensus algorithms). Most of these structures take the form of a Directed Acyclic Graph (DAG).
A DAG is a unidirectional data structure composed of a collection of vertices (squares) that are connected to each other by edges (arrows). In the context of DLT, the vertices represent transactions that have been added to the network’s ledger. The DAG model allows a transactional history to emerge as new transactions build on top of previous ones.
The tangle
The IOTA network uses a type of DAG structure called “the tangle”, in which a user has to approve two previous transactions before it can append a new transaction to the ledger. In theory, this network should have no transaction fees or processing delays, as there is no need to wait for transactions to be packaged into blocks. Instead, when users send new transactions to the network, they are at the same time confirming earlier transactions. As a result, IOTA offers support for micropayments and is thereby well-suited to the internet of things (IoT) and the machine-to-machine (M2M) economy.
However, it is too soon to tell whether IOTA is free from vulnerabilities or if it will be able to withstand cartel attacks. At the current stage of development, the IOTA network relies on a single node with special powers to reach consensus (The Coordinator), which is controlled by the IOTA Foundation. Every 2 minutes, The Coordinator sends a milestone transaction and all of the transactions that it links back to are considered to be confirmed. The IOTA team argues this centralised solution is necessary to protect the network during the early stages of development. An analogy can be drawn with Satoshi’s early mining in the Bitcoin network, which helped to make the network robust at the start. It is currently not known when the network will be robust enough to thrive without The Coordinator. Some have even argued (with admittedly little evidence) that the network will never be able to support itself without The Coordinator.
Hashgraph
Hashgraph is a DAG-based structure that was created by Leemon Baird from Swirlds in 2016. In the Hashgraph data structure, transactions are stored and transmitted in data packages called “events”, as opposed to blocks.
In blockchain-based networks, the consensus mechanism selects a node to create a new block of transactions. However, in Hashgraph all of the networked nodes communicate the latest state of their information to each other. In addition, their communications are communicated to arrive at a graph of exchanged communications that looks as follows.
Hashgraph’s consensus mechanism, known as Gossip about Gossip, is closely connected to the structure of this DAG. Standard Gossip protocols are used in all DLT networks to spread the news about transactional information. However, an additional consensus mechanism is required to establish an order of legitimate transactions.
In contrast, Gossip about Gossip goes one step further and enables nodes to communicate when and from whom they received new transactional information. This means that everyone knows exactly what everyone else knows. As such, Gossip about Gossip allows the nodes to reach consensus by simply examining the information that other nodes have relayed.
Thanks to the Gossip about Gossip algorithm, Hashgraph is highly scalable (limited only to the speed of the internet connection) and is thought to support up to 250,000 TPS in a permissioned environment. It is also computationally light and allows the chronological order of the transactions to be documented and guaranteed. This last point means that Hashgraph is the first asynchronous Byzantine Fault Tolerant (aBFT) system to be developed without compromising on scalability.
Hashgraph is patented by Leemon Baird and was initially designed to be implemented in permissioned networks. However, Hedera is the first public implementation of the Hashgraph data structure and the associated Gossip About Gossip consensus algorithm. In Hedera, users must pay transaction fees to nodes that process transactions, with 10% of the transaction fees going to Swirlds.
Hedera’s code is open review (as opposed to open source), which means that users may examine this and make suggestions, but may not use the code without a licence from Swirlds (except to build a wallet). In turn, the code can only be altered by agreement of the Hedera Governing Council, a group of 39 leading companies (with certain competing interests) that span 18 diverse industries and 6 continents. While software forks may still arise, nodes are encouraged to update their software to prevent their contributions from being deemed illegitimate.
Note: in the public implementation of Hashgraph, a Proxied Proof-of-Stake (PPoS) consensus mechanism will be used to reinforce Gossip about Gossip to prevent DDoS attacks. PPoS is a variant of DPoS where nodes can stake tokens against a specific node to increase that nodes weight in the voting process. Nodes in Hedera get paid interest in accordance with their accumulated stake, though they must share a proportion of their interest with those that staked on them.
Block-lattice
To make things even crazier, Nano came up with the block-lattice data structure, which is a hybrid of the blockchain and DAG structures.
Each account on Nano has its own blockchain, called an account-chain. Importantly, each account owner is in charge of updating their own account.
Note that light proof-of-work is used in IOTA and Nano when updating an account-chain to prevent DoS attacks (e.g. spam), as originally envisioned by the creators of this algorithm.
Conclusion
Congratulations on making it through! In this Lesson, we explored how the interconnectivity of DApps can help to give rise to the Distributed Economy. We also examined alternative forms of distributing tokens, various consensus mechanisms and novel DLT structures. Join our Telegram group for more developments :)
Consolidation exercises
- Examine some of the airdrops that are currently live and take a look at what the requirements are in order to receive tokens
2. Explore CoinList’s SAFT agreement, which is available at the link below
Extend your knowledge
