The structure of the DAG allows for greater scalability. Fantom's DAG 2.0 model uses graphs and event blocks that allow for multiple blocks of transactions to be confirmed simultaneously, as opposed to one block at a time (Ethereum).
The Lachesis protocol allows the network to achieve consensus using asynchronous Byzantine fault tolerance, or aBFT. In other words, it's similar to a network that adopts Byzantine fault tolerance, where even if one-third of the nodes are malicious, the network can still be trusted to validate and produce blocks with the correct order and timing.
The ‘asynchronous’ part means nodes can process and convey information at different times. As such, aBFT networks allow for some messages to be lost or indefinitely delayed. Although it's much more challenging to identify bad actors if messages can be delayed indefinitely, it showcases the network’s reliability and practicality in a more realistic setting.
As mentioned previously, many different networks and execution layers can be built on top of Fantom with Lachesis at its core to provide consensus and security. The first of these layers, a Layer-1 smart contract platform that's compatible with the Ethereum Virtual Machine (EVM) known as Opera, was launched on December 27th, 2019. This allows developers to create various decentralized applications or port them over from Ethereum or other EVM-compatible networks, such as Polygon or BINANCE Smart Chain.
Fantom can be distinguished from prominent PoW blockchains in a number of ways. Fantom is an example of Distributed LEDGER Technology (DLT) built on Directed Acyclic Graphs (DAGs). BFT blockchains such as Bitcoin or Ethereum are synchronous, meaning that there’s a lot of communication between nodes and consensus on the network’s state is achieved at the same time across all nodes. Asynchronous BFT (aBFT) consensus mechanisms, such as Fantom’s , stand to enjoy faster time-to-finality than BFT networks involving PoW.
Fantom’s consensus mechanism is also ‘leaderless,’ meaning no node has a special or privileged role in block production. is open-source software for networking between peers, using state replication to effectively enable multiple computers to function as one.
A Directed Acyclic Graph (DAG) is a mathematical abstraction for modeling nodes in a network (or a graph). Connections between nodes are called ‘edges’ in network theory, and these edges are directional in DAGs. DAGs are similar to trees, though trees are typically undirected (the connections between nodes aren’t one way or the other, they’re just connected) and don’t allow for closed loops. Trees have the interesting property of being both a graph and a data structure, which lends them naturally to modeling data on networks. DAGs have applications in scientific theory, practice and computation generally, ranging from epidemiology to scheduling.
In Lachesis, each node has a locally stored DAG of ‘event blocks,’ which describe transactions on the network along with their order. The chronology of events is elegantly described by directed graphs, enabling each node to independently (of the other nodes) determine a local view of the order of transactions. Each node uses its event blocks to produce final blocks for the network. Validator nodes then communicate with peers regarding transactions. Lachesis essentially involves peers agreeing on the root of DAGs, which ensures the ordering of events is agreed upon across the network.
Fantom’s blockchain doesn’t have a mempool of pending transactions as the network, as a whole, doesn’t agree on one block at a time. Nodes in Lachesis don't send one another blocks to confirm. Instead, the nodes pass around messages about events. Validators aren’t voting on blocks representing the truth across the entire network. Rather, they communicate regarding events and transactions. New events on the network are used as a means of voting on the events of the previous election cycles. Further, the DAGs in Lachesis are split into sub-DAGs called ‘epochs.’ Once an epoch reaches a sufficient number of blocks, it's ‘sealed.’ Fantom’s Lachesis protocol uses Lamport timestamps to achieve deterministic finality, unlike BFT blockchains such as Bitcoin and Ethereum, which are probabilistic (the exact time taken to generate and confirm sufficient blocks after a transaction varies). Where the Lamport timestamp is defined by the logical progression of directed edges in the graph rather than physical clocks.
Fantom Opera Chain
Fantom’s Lachesis protocol formulates the necessary DAG-based structure for its novel Opera Chain. This chain consists of nodes utilizing Fantom’s consensus model to create a chain of confirmed blocks. Fantom then stores data in these event blocks, which effectively represent data structures created by single nodes to share all information with the rest of the network. This is slightly different from traditional blocks in a blockchain.
Fantom is built on Scala, a high-level programming language that supports both functional and object-oriented programming. That's to say, choosing Scala gives Fantom a reasonable amount of technical flexibility. The Opera chain is designed to be EMV-compatible (the Ethereum Virtual Machine supporting smart contracts and dApp development), meaning applications can easily interact with data from Ethereum’s smart contracts written in Solidity or Vyper.
Each network on Fantom is independent of the others and functions as a network of networks, enhancing performance and reducing congestion. In addition, Fantom's mainnet uses EVM, making it interoperable with Ethereum. This compatibility enables developers to swiftly migrate to the new platform to take advantage of enhanced performance and cheaper costs.
Fantom is also designed to support the Cosmos SDK built atop their aBFT consensus mechanism. The Fantom Foundation is working with the University of Sydney to iteratively develop the Fantom Virtual Machine (FVM) (along with an associated database and interpreter) with the goal of improving Ethereum's security and performance.
As discussed in other sections, Fantom’s Opera chain leverages a Proof-of-Stake (PoS) consensus mechanism and features a leaderless aBFT approach to aligning the nodes in the network around one view of the truth.
The Opera Chain consists of three sets of components:
- High transactional throughput, theoretically in the hundreds of thousands of transactions per second (TPS); though this hasn’t been tested ‘in the wild’ yet
- Smart contract functionality via open source APIs
- Third-party application support across verticals
The diagram above shows how event blocks communicate with other blocks throughout the network. Each block in the chain stores the communicated information, creating a new event block. During this process, transactional information is hashed and passed along with each block containing hashed value of one or more of the preceding blocks in the chain. This is how Fantom achieves an immutable data structure.
Event blocks communicate with one another within the Opera Chain until they reach a Clotho block - a block containing a data structure called a flag table that saves all the connection data between different event blocks. This block must have a majority of more than ? connections with all previous event blocks. Clotho blocks then communicate and connect with other Clotho blocks.
Cloth blocks come to their own consensus to create another block called an Atropos block. Each Clotho block has a timestamp that represents when it was created. This information is passed along, and if at least ? of the nodes on the network are on the same timestamp, then the Clotho block becomes an Atropos block which are then strung together to actually formulate the main chain of the Fantom ecosystem.
The main chain of the system includes a total, timestamped order of all events on the blockchain. Participating nodes on the Fantom network carry a copy of this main chain, ensuring that they all have the ability to search for historical information within the Lachesis consensus protocol. This also allows for nodes to avoid storing the information of every single event block and, instead, reference the main chain.
