The Importance of Consensus

Consensus, as you know, means the method used to come to an agreement. So, consensus in a blockchain is how the blockchain nodes “come to agreement” over the blockchain transactions that will be written to the blockchain ledger. Consensus can be viewed as an agreement on the last state of the blockchain ledger's “world state,” which is similar to a snapshot or picture of the current transactions written. Consensus validates transactions and also orders them.

More specifically, a consensus algorithm is a process used to achieve an agreement, for example, for a transaction on a distributed network. The primary concern to the blockchain network's operation is the maintenance of the consensus of the information being recorded on the blockchain within the blockchain network. Consensus algorithms inherently have a trade-off. The trade-off is between transaction security and performance in most scenarios. Performance in blockchains is measured in transaction throughput, which is also known as transactions per second (TPS). For example, the nature of Ethereum is trustless and is addressed by using the well-known proof-of-work algorithm. The Ethash consensus algorithm used in Ethereum makes attacks both prohibitively expensive and unlikely to occur. However, Ethereum is slow compared to other database technologies.

Consensus can also impact the parameters and security of the blockchain ledger operations. Understanding the strengths and weaknesses of the blockchain consensus being deployed is advisable since exposure to known vulnerabilities can be avoided with some basic knowledge. Also, the blockchain network could have rogue actors that could facilitate greater exposure to vulnerabilities.

Enterprises require availability and consistency at a minimum. By the end of this chapter, it should be clear that specific consensus algorithms were designed to achieve reliability in a network involving multiple unreliable nodes. Solving the consensus problem as it is known in the industry is quite important in distributed computing and for that matter enterprise services.

Consensus in a blockchain follows the same requirements as distributed computing and must satisfy the following two properties to guarantee an agreement among network nodes:

• Safety is referenced in most blockchain platforms as being able to provide a finality to a blockchain transaction. In a blockchain, this can mean that each node will have the same output for each input.
• Liveness is referenced to availability. In a blockchain, this means that each nonfaulty node will eventually receive every submitted transaction.

Two types of blockchain consensus are generally accepted in the industry: voting-based and lottery-based. Voting-based consensus should be chosen based on the following:

• Finality—Voting-based algorithms are considered beneficial since they provide what is considered low-latency finality. This is accomplished by a majority of nodes validating a transaction or block.
• Strict rules—Rules are strictly enforced, especially in a PoW consensus method as used in Ethereum.

Lottery-based consensus should be chosen based on the following:

• Fairness—The function should distribute leader election across the broadest possible population of participants.
• Investment—The cost of controlling the leader election process should be proportional to the value gained from it.
• Verification—It should be relatively simple for all participants to verify that the leader was legitimately selected.

Byzantine Generals Problem

The Byzantine Generals Problem comes from the world of traditional computer science. In this scenario, the involved parties must come to an agreement for a strategy in order to avoid a complete failure of operations. It could also be that some nodes in the network could be corrupt and effectively spreading unreliable requests or information.

Before telecommunications, encryption, and general information technology (IT), the only way to really communicate with other people was through other people who were messengers. This form of communication was clearly dangerous for both the army and the messenger. Messengers could be captured, for example, and the message stolen, which could place the army in a vulnerable scenario. Sure, other forms of communications were available such as smoke signals or mirroring. However, smoke signals were not accepted as a way to be secretive.

Perhaps the city is strong enough to defend itself against one or two of the enemy army brigades, but it may not strong enough to defend against three and definitely not against seven. So, the general's seven brigades need to have “consensus” and must agree on how, when, why, and where to attack with precision.

How do the generals attack at the same time? How do they know a message sent from another brigade was not tampered with? How do they communicate so it's not intercepted?

Now let's get back to computer science and specifically discuss how consensus affects blockchain technology. In the world of blockchains, nodes are essentially virtual machines running the blockchain networking protocols, code, and messaging services on a distributed network. These blockchain nodes need a way to reach an agreement when it comes to writing to the blockchain ledger. This is where consensus comes in to handle how these blockchain nodes come to an agreement in this distributed network.

Figure 4.1 shows a blockchain network with Ethereum Virtual Machines. This distributed network has six nodes connected in a mesh network style. Blockchain ledgers are not updated on one node at a time but are actually propagated to all the nodes at the same time in most distributed ledger platforms.

It is critical that the agreement between all of these blockchain nodes on how to write to the blockchain ledger is strictly defined. This strictly defined agreement in blockchain is called reaching a consensus.

The true solution to the Byzantine Generals Problem is not a straightforward solution where “one size fits all” in the world of consensus. These blockchain ledger solutions need to involve specific types of hashing, intense computing work, and a latency-tolerant peer-to-peer communications protocol between all the nodes to verify the transactions. You can think of nodes as generals when applying the Byzantine Generals Problem to blockchains.

Byzantine Fault Tolerance

Byzantine fault tolerance (BFT) came about since it represents a valid solution to the Byzantine Generals Problem. BFT is a crucial part of an effective blockchain platform, and there are multiple ways in which tolerance can be implemented. In your role as a presale's specialist or IT architect, you need to understand the various consensus methods available based on the enterprise blockchain platforms you're considering.

Figure 4.2 goes through the step by step of how the Byzantine Generals Problem works. First, General A gets a message to attack, while General B does not get this message. Because General B does get this message, this creates a problem where the city that was being attacked could foil the whole attack.

I will be discussing various forms of Byzantine fault tolerance, such as PBFT and other forms of BFT, in the remaining parts of this chapter.

Proof-of-Work Consensus

Bitcoin was the first practical and successful cryptocurrency platform, and it introduced the proof-of-work (PoW) consensus method as part of the platform. The proof-of-work protocol involves block miners solving complex cryptographic puzzles. As part of participating in the challenge of solving complex problems, miners are compensated by receiving rewards in the form of Bitcoin in the case of the Bitcoin blockchain platform, or Ether in the case of the Ethereum blockchain platform. Proof of work is like a running a marathon for the blockchain miners in the sense that the first node to produce the longest chain will win the block rewards, which are cryptocurrency or tokens depending on the platform.

Proof of work is the most widely used consensus method. This is mainly because it was the original protocol and has proven its resilience against internal and external attacks. Basically, PoW demonstrates that a participant has done some work and gets a reward for solving a problem.

A high-level overview of the PoW consensus in Bitcoin is that a block including relevant parts of the transaction is hashed, and a random nonce is added to it so that the resulting hash is below a certain value, which is called the difficulty level. Another way to consider the mining process is to look at it as an operation of inverse hashing, which is a cryptographic approach. This inverse hashing determines a random number (nonce) so the cryptographic hash algorithm of block data results in less than a given threshold, in other words, the difficulty level. This difficulty level is what the miners use to gauge the level of compute power and resources needed to mine Bitcoin, for example, and if it would be profitable.

Figure 4.3 shows the high-level process of a PoW transaction where it references the previous block's hash.

The second block is referencing the first block's hash, and then the third block references the second block's hash. For a block to be written to a blockchain, significant work has to be done by the miners.

Proof of work's main characteristics are the following:

• It provides hard-to-predict blocks, which can add to the competitiveness of the blockchain for miners but also security for the blockchain network. Essentially, the math problem gets more difficult as required by the difficultly rate.
• It provides an easy way to validate the correctness of blocks by validating who manages to solve the problem through predefined values.

One main challenge with PoW is that a lot of energy is considered “wasted” or “consumed.” When you have thousands of nodes on a network such as Bitcoin, you will have thousands of nodes working to solve a problem. Consumption of resources is the main concern. Along with that, the cost to mine a Bitcoin has skyrocketed as well to over $1,000 or more. Mining cryptocurrencies requires intense electrical power requirements. To successfully mine Bitcoin, for example, you would need to purchase a “mining rig” that would likely have an application-specific integrated circuit (ASIC), which is a microchip designed for a special application such as Bitcoin mining. Mining is a competitive business, and an ASIC can provide a competitive edge and provide efficiency, performance, and even security in the blockchain mining arenas.

For example, in Bitcoin, a miner needs to be competitive, and this is done by continuing the testing of different unique values. These unique values are known as nonces. Once a miner manages to solve this complex problem, the miner will receive the prevalent bitcoin rewards. The miner then may add the block into the Bitcoin blockchain by broadcasting to the blockchain network that the block has been mined by this node.

Mining serves two main purposes for a blockchain. First, mining is used to verify the legitimacy of a transaction and because of this, helps to avoid the so-called double spending issue. Second, mining facilitates the creation of rewards and disbursement to the miners on the Bitcoin network.

If you're an enterprise or your customers are enterprises, then you're likely not going to be using a PoW-based blockchain such as Bitcoin. You may want to use Ethereum, for example, as a token platform or perhaps to extend a payment gateway for a cryptocurrency platform as well.

In the case of Quorum and Enterprise Ethereum, however, you may have a good use case to consider those blockchains as well because of their hybrid enterprise solutions. Chapter 8, “Enterprise Blockchain Use Cases,” covers the more common use cases for Quorum and Enterprise Ethereum.

Proof-of-Stake Consensus

Proof of stake (PoS) is a consensus algorithm that is commonly used by cryptocurrencies to validate blocks and is very different from proof of work from a consensus standpoint. Proof of stake was created in 2011, and the first cryptocurrency to implement it was Peercoin in 2012. Essentially it was created as a way of avoiding the well-known economic and environmental issues with PoW, such as intense energy consumption and cost of mining.

Proof of stake implements an approach where the creator of the next block is determined by a randomized system that is logically dictated by how much of an investment a user is holding or how long they have been holding that particular currency. This is a different approach from the “computational power” in PoW, for example, where the probability of creating a block and receiving the associated rewards is proportional to a user holding the underlining token or cryptocurrency on the network.

Proof of stake is an interest-based approach to handling a blockchain where a node's interest is directly proportional to its investment. Another way to look at proof of stake is that it is a deterministic way that is essentially based on wealth, defined as stake. A greater investment in a cryptocurrency stake equates to greater influence.

Proof of stake's main characteristics are as follows:

• A stake may change due to economics.
• Votes are based on economics.
• Nodes need to be online to vote.
• Votes are final.
• Multiple votes are not allowed.

Ethereum has a planned Casper release, which is a fork of the Ethereum blockchain. This fork would be to change from a PoW consensus to a PoS consensus. Note that if Ethereum moved to a PoS consensus, it would be a major disruptive move, since Ethereum would go from a mining consensus to a validator consensus.

The PoS system is well enabled for platforms with a static coin supply. For example, most crowd-funded platforms leverage this approach to distributing tokens based on investment. This is exactly where Ethereum is going. However, what makes Ethereum different is its PoS-based finality system capable of overlaying an existing PoW blockchain. This overlay on top of PoW is essentially a hybrid PoW/PoS approach called Casper Friendly Finality Gadget (FFG). However, this update has been delayed because of technical challenges and security concerns.

One of the main advantages of using proof of stake is that it has a significantly lower energy requirement; therefore, you can get a better return on investment (ROI).

It has been documented that each Bitcoin transaction, which uses a PoW system, can require as much electricity as an average Dutch household does in two weeks. This is both ineffective and unsustainable.

Proof of stake is considered a more efficient consensus protocol as it requires far less electricity to operate and can run on less strict hardware requirements such as ASICs instead of GPUs.

Also, PoS has less of a need to release many new coins, which has been a means of incentivizing miners to maintain the network. This helps keep the price of a particular coin more stable and therefore provides incentive for more blockchain participation.

Comparing Proof of Work and Proof of Stake

Now that you have a general understanding of both PoW and PoS, let's review PoW versus PoS to understand how disruptive it would be for Ethereum to move from PoW to PoS.

• Proof of work—PoW relies on miners running nodes on the network to solve computationally difficult math problems to validate new blocks of a transaction. Miners are compensated for their work.
• Proof of stake—PoW relies on validator nodes on the network to take turns proposing and validating the next block in the chain. The value of the validator's node—and the size of its reward—depends on the amount of coins staked in the verification process. Essentially, the more you deposit, the bigger the potential return.

PoW is work intensive and requires miners (nodes) to write to the blockchain transactions. Miners are costly, and the mining process is hideously intensive economically and from an energy standpoint.

According to a recent Cointelegraph article (https://cointelegraph.com/news/bitcoin-mining-uses-more-power-than-most-african-countries), it was estimated that just the Bitcoin network uses 0.14 percent of the global energy consumption. Bitcoin mining now potentially consumes more electricity than the bottom 750 million electricity users, which is more than 10 percent of the population of the world.

PoS requires a different type of investment than PoW and is less energy intensive.

PoW offers the following pros:

• It is a historically proven consensus method that is widely used for cryptocurrency and is a stable platform.
• It's secure because of the high cost and low probability of a 51 percent attack. (A 51 percent attack is where the pool of mining nodes is taken over by a group of miners controlling more than 50 percent of the network's mining hash rate.) This type of attack is unlikely because of the immense hash power and the requirement for the number of nodes to control.

PoW has the following cons:

• Transaction processing is slow compared to other consensus approaches. For example, in Bitcoin it can take several hours to get a transaction confirmation.
• The 51 percent attack concern is still real, meaning that 51 percent or more than 51 percent of nodes in the network can be exploited.
• Mining is a time-consuming process because of the work that has to be done to produce blocks.
• Mining is an expensive process from both an economic and environmental perspective. For example, Bitcoin mining uses so much power it is frequently banned by municipal power companies.

PoS offers the following pros:

• It is energy efficient compared to proof of work since there is no need to mine coins, making it an environmentally friendly option to cryptocurrencies such as Bitcoin.
• A PoS system has a lower barrier to entry because no mining rigs are required, and it allows more users to take part in the staking and forging processes.

PoS offers the following cons:

• Someone can monopolize the network. With PoS you know that an increased stake in the network means increased leverage. This can be a concern if someone forges most of the future blocks and centralizes the rewards.
• Hackers can steal your wealth but also your stake in the network. When someone loses their “wallet” with the coins, then they also lose their ability to vote in the blockchain network.

As with any blockchain requirement you're trying to establish, reviewing the benefits and drawbacks should enable you to determine the right solution.

Proof of Elapsed Time

Proof of elapsed time (PoET) attempts to address the problem of proof of stake where the random election of participants proposing blocks can occur and also ensures that every participant has a fair chance to propose a new block and then participate in the voting process.

PoET was developed by well-known computer chipmaker Intel to be a production-grade protocol capable of supporting large network use cases. PoET requires a special CPU instruction set called Intel Software Guard Extensions (SGX), which provides for a trusted environment called TEE. Trusted code runs in an environment that is private from the rest of the application. This means the rest of the application will not interfere with the memory space of the trusted code, for example. Think of this as a container in cloud computing or a sandbox.

PoET essentially mitigates concerns about the PoW consensus by electing what is considered a leader and implementing a two-tier process. PoET attempts to correct the issue of PoS, which will arbitrarily determine that the members proposing blocks are expected to guarantee that each member has a reasonable opportunity to offer a block.

Every participant in the network is assigned a random amount of time to wait, and the first participant to finish its wait time will have the opportunity to commit the next block to the blockchain.

PoET imposes a hold-up time from its local reliable enclave of node members. The node member with the shortest hold-up time is next to “offer” a block only after the expiration of the hold time. Each privately trusted enclave signs the potential requests. The results are then validated by other members so they can confirm that no other nodes have skipped the waiting time.

There are valid concerns that because PoET relies heavily on SGX for the foundation of the protocol, the SGX enclave could be hijacked. This has been proven with the recent Foreshadow vulnerability. Still, Hyperledger Sawtooth utilizes the PoET consensus algorithm, which leverages Intel's SGX to implement this leader-based lottery system.

PoET is a specialized use case. Its main benefit is that it is a trusted proprietary platform from Intel.

PoET offers the following pros:

• Trusted environment that is essentially a plug-and-play consensus on Hyperledger Sawtooth
• Lower energy usage than a PoW consensus that has mining costs such as hardware and energy
• Solves the random leader selection problem without being resource intensive or requiring incentives

PoET has the following cons:

• Lack of portability due to reliance on Intel's proprietary hardware and software requirements
• Higher costs than other blockchain platforms due to the proprietary hardware requirements

As you can see, the pros and cons could certainly be acceptable for enterprises that are seeking a targeted solution and not expecting to go outside of that solution.

Delegated Proof of Stake

Delegated proof of stake (DPoS) was invented by Daniel Larimer, cofounder of Steem and CTO of EOS, both of which use DPos. DPos is an offshoot of its relative PoS.

DPoS uses a real-time voting system and a reputational system to achieve consensus. The DPoS blockchain consensus protocol allows the blockchain token holders to leverage their coin balances to elect delegates that are called witnesses. These witnesses have the opportunity to stake blocks of new transactions and then will be authorized to add them to the blockchain network.

Interestingly, the voting power is determined by how large the token holdings are of the specified blockchain stakeholders. The stakeholders who have more coins or tokens, for example, will have a greater impact on the network than those with fewer. Generally, this impact of the stake is directly proportional to the stake that has been placed into the system's network.

DPoS is considered to be the least centralized consensus protocol compared to all others as it is the most inclusive.

There are significant variations of DPoS where a delegate needs to show commitment by depositing funds into what is similar to a time-locked security account. Each of these blockchains also has a different protocol for how the consensus method is implemented and maintained.

There are typically 21 to 101 delegates elected in the various blockchain networks using the DPoS consensus algorithm. For example, at the time of writing, in EOS and also Steemit there are 21 block producers. In Bitshares, there are 101 block producers. In a DPoS network, it is up to the consensus rules of that chain to determine the variables of choosing a block producer.

Essentially, there are several benefits such as availability and performance, which are just two differentiators as compared to other blockchain algorithms. These are clear benefits to using a DPoS consensus.

DPoS offers the following pros:

• DPoS consumes significantly less energy than PoW and is considered to be very energy efficient.
• DPos provides incentives to participate in the network and provides delegates with a way to be voted out if they misbehave on the network.
• DPos scales greater than PoW and PoS and has faster transaction processing due to less overhead.
• There is a fair reward distribution where it is considered a “democratized” reward schema.

DPoS has the following cons:

• A 51 percent attack is clearly possible because fewer people are in charge of ensuring the network stays secure.
• Centralization of control could occur if the delegates with the most tokens gain strength, which in turn could allow a “cartel” approach to network control.

Some of the blockchain projects that use a DPoS consensus include Steemit, EOS, Bitshares, and Lisk.

Delegated Byzantine Fault Tolerance

Delegated Byzantine fault tolerance (dBFT) was developed by the NEO team to overcome the Byzantine Generals Problem. NEO is similar in many respects to Ethereum, but I chose not to cover NEO as one of the blockchains due to its low usage and unproven platform at the time of writing.

It is, however, important to note that NEO does have some features that would be useful for enterprises, so covering it here will be useful.

The system comprises nodes, delegates, and a speaker. dBFT essentially works in a similar fashion to a country's government structure, with its citizens, delegates (representatives), and speakers (politicians) to ensure that the country (network) is functionally correct.

dBFT has some unique terms compared to other blockchain consensus, so it's important to clarify some terminology before moving on.

• Citizens are essentially NEO token holders and are considered ordinary nodes. A token holder is anyone who holds a cryptocurrency token and maintains an interest in the blockchain by voting.
• Delegates are bookkeeping nodes and elected to the role to file requests.
• Speakers are randomly chosen delegates to follow the citizens' requests by proposing the requests.

dBFT is a unique consensus algorithm developed for NEO with what can be considered a perfect finality. Perfect finality means that all transactions are 100 percent confirmed to be final after the first confirmation.

Interestingly, the blockchain cannot fork with dBFT, and this can remove some friction for stakeholders. dBFT is totally focused on the enterprise since it was clearly built with both regulatory and business use cases in mind.

dBFT offers the following benefits:

• NEO offers immediate finality after confirmation as well as being an almost impossible protocol to launch a 51 percent attack.
• NEO has fast and efficient protocols since a new block on the chain takes between 15 and 20 seconds.
• The NEO dBFT network cannot be forked and, therefore provides stability for participants.

dBFT has the following disadvantages:

• There is high centralization due to the low node count of the NEO network.
• Centralization is clear since the bookkeepers are controlled by the NEO council.
• The NEO network cannot actually be forked, which may be a concern if a disagreement occurs between members.
• The NEO user and developer bases are still small compared to Ethereum, which makes expertise harder to find.

NEO is a nonprofit community-driven blockchain project. It utilizes blockchain technology and digital identity to digitize assets and automate the management of digital assets using smart contracts. To find out more about NEO and dBFT, go to https://neo.org/.

Practical Byzantine Fault Tolerance

Practical Byzantine fault tolerance (PBFT) has been the most widely used permissioned blockchain platform protocol at the time of writing. PBFT was introduced by Miguel Castro and Barbara Liskov at the MIT Laboratory for Computer Science in 1999. PBFT is also one of the several potential solutions to the Byzantine Generals Problem, which was discussed earlier in the chapter.

PBFT consensus decisions are determined based on the total decisions submitted by all the generals (nodes). PBFT addresses the challenges without the extensive expenditure of energy required by proof of work. It is important to note that PBFT works only on a permissioned blockchain and thus does not allow anonymity like in Ethereum.

PBFT has the main purpose of deciding whether to accept a piece of information submitted to the blockchain. Essentially, is the information being proposed from an honest and reliable source?

PBFT offers the following two benefits:

• Rapid transaction finality that does not wait for confirmations to log transactions
• Lower power consumption compared to PoW

PBFT has the following disadvantages:

• Initially designed for a limited use case because of the high load of network traffic between nodes
• Can be susceptible to sybil attacks where a single party can create or manipulate a large number of network nodes, compromising the network

Blockchain projects such as Zilliqa, Hyperledger Fabric, and Ripple currently are using PBFT as their primary algorithm or have it as a choice in their portfolio.

Istanbul Byzantine Fault Tolerance

Istanbul Byzantine fault tolerance (IBFT) is a hybrid form of BFT and is an efficient alternative to PoW, which is currently used in the Ethereum network. IBFT is an implementation of the PBFT algorithm with some significant modifications to the blockchain code. These modifications provide for benefits that include efficient settlement finality and reduced infrastructure.

For example, in its use case with Quorum, IBFT uses a pool of validating nodes operating on a private Ethereum network, which will determine whether a proposed block is valid to be proposed for addition to the blockchain network.

The next step in the process that occurs is that one validating node is arbitrarily selected as the proposer and will be responsible for constructing a block at the block interval. It then will share this block with the group.

Effectively, if a “super-majority” of the node validators accept the block to be valid, the accepted block is then written to the blockchain.

Lastly, at the completion of the consensus process, these validator nodes may select a new proposer. This proposer then will be responsible for providing the next candidate block at the next block interval in the process.

IBFT is a Byzantine fault-tolerant solution offering immediate transaction finality that reduces the required blockchain infrastructure that other platforms may require. IBFT offers substantial benefits when used on a private blockchain where the validator pool is trusted and held accountable.

Lastly, IBFT really provides for what is a predictable transaction processing rate that enterprise blockchain must have.

The most significant implementation of IBFT is in the Quorum blockchain.

Raft Consensus

Raft is a consensus algorithm that is designed to be easy to understand by most IT professionals in the sense that it has a simple ledger structure with little overhead. The main difference is that Raft has been simplified into relatively independent processes, which are known as “subproblems.”

The Raft processes are as follows:

• Raft provides for a leader election process, where a new leader is elected in case of the failure of an existing one.
• Raft provides for a log replication service for the leaders, which provides high availability.
• Safety is the process that is implemented if one of the servers has committed a log entry at a particular index and, comparatively speaking, no other server can apply a different log entry for that specific index.

Raft has provided several useful resources to help you learn Raft efficiently. For an interesting and interactive Raft visualization that you can view in your browser, visit https://raft.github.io/.

For more information about distributed consensus, the leader election process, and Raft processing, visit http://thesecretlivesofdata.com/raft/.

Figure 4.4 provides a visual of the Raft processing resource. As you can see, there are five servers in a Raft cluster that is running in a browser. What is interesting is that you can interact with the utility.

Raft offers the following benefits:

• Raft has a very simple ledger and network structure that you can comprehend quickly.
• Raft has been well known since it is a “sister” version of Paxos. Raft is used as a backend data structure for containers.
• Raft uses a randomized election timeout process that ensures that “split votes” are minimized.
• Raft consensus does not mint blocks unless there are pending transactions, which can provide for some efficiency.
• Raft transactions have a faster block time as compared to IBFT consensus used in other blockchains.

Raft has the following disadvantages:

• Raft has a somewhat limited use case in the sense that it is not used directly with cryptocurrencies.
• Raft also has a limited enterprise blockchain ledger that is available to work with significant enterprise use cases.

The Raft consensus is an option used in the Quorum blockchain

Directed Acyclic Graph

Directed acyclic graphs (DAGs) may become more prevalent in ledger technology's future based on several factors. Some of these factors are greater performance and significant scalability, which in reality blockchains do not do well.

A DAG is a very different data structure than what a blockchain is. A DAG has traditionally been used in computer science to solve challenges around data modeling and data analysis.

Before moving on, it is important to understand some definitions when discussing the DAG platform. DAG's have some unique terms compared to other blockchain consensus methods.

The following terms are unique to DAGs:

• A web is a network consisting of nodes connected to each other with edges.
• An edge is a one-way connection point between one or more network nodes.
• Acyclic means that a transaction cannot encounter the same node for the second time when moving from one node to another node by following the edges of the network.

Generally, it can take 10 minutes to create one block, and it's important to note that blocks cannot be created simultaneously. What is actually significant is that transactions can be running on different chains simultaneously, and this can provide better performance.

What is the difference between a blockchain and a DAG?

Blockchains are a very different data structure where a distributed ledger forms a linear chain of blocks of transactions in an immutable chronological order that is clearly timestamped. A DAG is effectively different since it is a network of individual transactions linked to any multiple other transactions.

A DAG does not transact serially like a blockchain rather in a form of a parallel ledger data structure.

A blockchain is a linked list of blocks, but a DAG is a tree where transactions branch out from one transaction to another. This is considered a graph that travels in one direction without cycles connecting the other edges. A simpler way to look at a DAG is a “web of network nodes” where each network node is interconnected but is running one-way communications to the other network nodes.

Figure 4.5 compares visually a typical blockchain structure to a DAG structure in Hedora Hashgraph. The structure of a blockchain is more linear and hierarchal as well. One of the main benefits of DAGs are that these networks are faster because of how the transactions are validated and then processed in a parallel approach. In a blockchain, transactions are processed in a serial approach.

DAGs offer the following advantages over a blockchain:

• DAGs are fast and can scale with more transactions.
• Transactions can be validated in a parallel or simultaneous manner, compared to a blockchain that validates in a linear fashion.
• Lower transaction fees may be realized because of efficiencies in protocols.

DAGs have the following disadvantages over a blockchain:

• Considered complex to learn from a historical perspective
• Limited uses cases where blockchains may not be appropriate such as with Internet of Things (IoT) data

Ledger platforms that use DAGs are IOTA, Hashgraph, and Nano.