Impact of System Design Features on Blockchain Use Cases

Contents

1. Introduction
1.1. Objective

2. Core Concepts Of The blockchain
2.1. Distinction Distributed Ledger Technology and Blockchain
2.2. Evolution ofBlockchain
2.3 Merkle Trees and Hashes

3. Features / Components
3.1. Public / Federated / Private Blockchains
3.1.1. Public Blockchains
3.1.2. Consortium Blockchains
3.1.3. Private Blockchains
3.2. Permissioned vs. Unpermissioned Blockchains
3.3. Consensus Mechanisms
3.3.1 The Byzantine General Problem
3.3.2. Proof-of-Work
3.3.3. Proof-Of-Stake
3.3.4. Delegated Proof-of-Stake
3.3.5. Proof-of-Authority
3.3.6. Proof-of-Burn
3.3.7. Proof-of-Importance
3.3.8. Proof-of-Elapsed-Time
3.3.9. Proof-of-Activity
3.3.10 Proof-of-Capacity
3.3.11. Federated Byzantine Agreement
3.3.12. Practical Byzantine Fault Tolerance
3.4. Smart contracts
3.5. Decentralized Applications

4. Blockchain Use Cases
4.1. Currency
4.2. Digital Identity
4.3. Supply Chain Management
4.4. Digital Voting
4.5. Health Data Management
4.6. Sharing Economy

5. Conclusion

References

ListofFigures:

1. Introduction

Information technology has already impacted business and industry practices in a profound way. With the rise of the Internet the world industries can interact easily with each other.

A new technology that recently has gotten a lot of attention is the blockchain - the technology behind Bitcoin and most other crypto­currencies. There is a big hype (and potentially bubble) about the crypto- currencys market which draws a lot of attention to the blockchain technology. But the blockchain is much more than just the technology behind bitcoin.

It lays the foundation for a new economic system, the “crypto-economy”. It is “not defined by geographic location, political structure, or legal system, but which uses cryptographic techniques to constrain behaviour in place of using trusted third parties” (Babitt and Dietz 2014). Cryptoeconomics is the practical science, that focuses on the design of the protocols, that make cryptoeconomy possible. Cryptography is used to prove prove properties established in the past (e.g. account balances, identities, ownership). Economic incentives defined inside the system encourage desired properties to hold into the future (e.g. Bitcoin mining) (Buterin 2017).

The blockchain is a “general purpose technology” (McPhee and Ljutic 2017). It can lead to “the creation of many sub-inventions” (Gordon, 2017) and therefore many argue that it is rather a “foundational” technology than a “disruptive” technology, as it is often argued in many papers on this subject. Therefore it might take many decades for the economy to implement the technology fully (Iansati and Lakhani 2017). The most potentially disruptive feature of the blockchain technology is the decentralization of transactions within the network. This means that all the transaction data is stored and encrypted by many participants in the network, instead of being managed by a single entity This way the blockchain is capable of removing the intermediary (centralized entity

who owns all data), which is almost always necessary in nowadays economy. Blockchains are called “trustless”, as participants do not need to trust each other or a third party with personal data for making safe transactions (McPhee and Ljutic 2017). In blockchains the intermediary is replaced by a cryptographic proof (Perfall et al. 2016). No matter what is bought or sold; it is most likely that there is some instance in between the vendor and the seller. As nodes interact individually with each other, without any instance in between them in a blockchain Peer-to-Peer network there is great potential for cutting transaction costs and also search costs. There is a huge potential to change the way individuals will interact if blockchain technology will be more widely adopted in the future. - If nodes become legal identities there is a huge potential for dismantling bureaucracy through the use of smart contracts.

Usual Transactionmodel

Abbildung in dieser Leseprobe nicht enthalten

Abbildung in dieser Leseprobe nicht enthalten

Fig. 1: Traditional Transactionmodel compared to blockchain Transactionmodel.

In blockchain networks transactions are stored and encrypted on the distributed ledger which is updated by consensus amongst participants. Smart contracts are individually defined rules (e.g. about prices and quantities) demand meet supplly autonomously.

1.1. Objective

There are many different ways to design a blockchain network. Depending on the specific use case, some options might work better than others. In this thesis, I am going to identify and distinguish the main system design features a blockchain consists of, to then show how they impact the functionality of the network and how it can be applied to different use cases.

Abbildung in dieser Leseprobe nicht enthalten

Fig. 2: The four-layer architecture of distributed ledgers. Adapted and modified from: Nadine Ruckeshauser (2017)

The architecture of distributed ledgers / blockchains can be split into four different layers. The separation allows the identification of layer-specific technological features and exposes the individual system design decisions, that have to be made at each layer when building a blockchain. (Ruckeshauser 2017)

Although the layers can be looked at separately, they are all inter-related e.g. decisions at the first layer affect the decision-making process at the second layer. This means, that the all-encompassing goal of the blockchain network should be determined before deciding on certain technological features. The overarching effects of each design decision must be taken into account in order to arrive at a blockchain solution that provides the desired properties.

2. Core Concepts Of The blockchain

2.1. Distinction Distributed Ledger Technology and Blockchain

The expressions “Distributed ledger technology” (DLT) and “blockchain” technology are often used synonymously. A distributed ledger is an append only tamper-resistent database that is not maintained by a central authority (Swan 2017, TIMReview). It is spread across several nodes, where each node has an identical copy of the ledger and updates independently. Updates require a consensus amongst the nodes. Once it is reached the database is updated to the the latest, agreed-upon version. R3's Corda from the financial sector is probably the best known distributed ledger.

The blockchain is a type of distributed ledger. The transaction data of a blockchain is organized in blocks. In the Bitcoin network these are created every ten minutes and cryptographically secured. Each block has a link that points to the previous block, hence all blocks are linked together forming a chain ofblocks - the reason for the name “blockchain”.

2.2. Evolution of Blockchain

Blockchain applications can be put into three distinct development stages. (Perfall 2016)

Blockchain 1.0 refers to crypto-currencies like Bitcoin, that could be an alternative to conventional currencies. It is the first and easiest application of the blockchain. Even though the number of users and transactions grows, the Bitcoins margin in the international currency-market is still really low (Perfall 2016).

Blockchain 2.0 is the next evolutionary step of blockchain technology which involves so-called smart contracts. Smart-contracts allow for automatic transactions after preset conditions of a contract occurred without involving a third party (Unibright.io 2018).

Blockchain 3.0 is all about dApps - short for decentralized applications. They avoid centralized infrastracture, storage and communication runs on a decentralized peer-to-peer network. It allows for non-financial uses and can be applied to optimize processes in bureaucracy, the health sector and others (Bouillon, M 2017).

2.3 Merkle Trees and Hashes

When the Bitcoin White-Paper: "Bitcoin: A Peer-to-Peer Electronic Cash System" was published in 2008 by Satoshi Nakamoto, an anonymous person or group of hackers, the concept of blockchain was first introduced. It is a distributed ledger / data-base where transaction data is time-stamped, encrypted and then put together into blocks where each block points to the previous block, so that you have a record of every transaction ever made. As all the blocks are linked together in a chain it is called blockchain. Bitcoin is a so-called cryptocurrency. Contrary to conventional fiat-currencys like Euro or US-Dollar it is notmaintained by a central bank, but runs on a decentralized network (distributed ledger) which is publicly available and lets anybody participate.

The first block in the chain is called the genesis block; it does not have a previous block and is usually hardcoded into the protocol. New blocks are discovered according the specified set of rules set by the cryptocurrency protocol. An important function of these rules is to protect against attacks on the blockchain and to reach consensus in case multiple instances of the blockchain appear. The repeated "hashing" of transaction data is essential for blockchains. The transaction data from the individual transaction is hashed by using a hash function that creates a single string out of all the transaction data. Then the hash value is hashed again with the hash of the next transaction. The process is repeated numerous times and by that a structure called a "merkle tree" is created. The "merkle root" is put into the block that gets added to the blockchain. It is a hash that consists of hashes of every transaction of the entire block. As soon as the block gets confirmed by the network it is time stamped and then added to the blockchain.

Abbildung in dieser Leseprobe nicht enthalten

Fig. 3: An exemplary blockchain and merkle tree graphically illustrated. Block 27 points to the previous block by having stored the matching hash value. The matching nonce for the merkle root has to be found by miners so that the block can get added to blockchain.

3. Features I Components

3.1. Public / Federated / Private Blockchains

There are several kinds of blockchains: Public, federated and private ones that differ in who has access to it, who is allowed to write in it and who controls the consensus protocol.

3.1.1. Public Blockchains

Everyone with internet connection can participate in a public blockchain. It is public as anyone can participate for free and unconditionally. No one is in charge or owns it. This evokes the question how decisions are being made in this type of network. They are made by the users, who validate transactions within the network to then update the public, decentralized ledger. Anyone can take part in the consensus process to determine what blocks get added to the chain and what the current state is. As public blockchains are considered to be “fully decentralized” they use cryptoeconomics to substitute for an entity of centralized trust - “the combination of economic incentives and cryptographic verification using mechanisms such as proof of work or proof of stake, following a general principle that the degree to which someone can have an influence in the consensus process is proportional to the quantity of economic resources that they can bring to bear" (Buterin 2015)

Public blockchains have several drawbacks. For one a substantial amount of computing power is necessary to maintain the distributed ledger at a large scale as consensus in public blockchains is usually reached by the Proof-of-Work algorithm, where a complex mathematical puzzle has to be solved in order to validate a block and to stay in sync with the latest version of the distributed ledger. This process is really ressource­intensive, slow and needs a lot of electricity. It is proven, that a single transaction on the most popular public blockchain Bitcoin costs approximately as much electricity as an American household uses within an entire month. The bitcoin network as a whole consumes almost as much energy as the Czech Republic and more (BitcoinEnergyConsumption.com). When comparing the electricity consumption of Bitcoin transactions to that of financial transactions made through the payment system VISA the difference is horrendous. Half a million VISA transactions cost as much electricity as a single Bitcoin transaction. A solution to this problem would a switch to a different consensus algorithm such as proof-of-stake. Another option is to guide the computing power towards useful scientifical research as GridCoin does it (Gridcoin.us). Public blockchains can fall victim of the so called 51%- attack. As the validators of it are unknown, a collusion of 51% of the miners could possibly take control of a public blockchain. Another disadvantage of public blockchains is the openness of it, anybody can check the transaction logs which provides little privacy for users of public blockchains. On top of that, transactions are irreversible as public blockchains are usually immutable. Irreversibility might be really useful in certain use cases such as land registries (Buterin 2015).

3.1.1. Consortium Blockchains

Consortium blockchains are maintained by a group of companies or a federation/consortium. They are mostly used in the banking sector and R3's Corda blockchain is the most famous example of a federated blockchain. Only members of the federation are allowed to run a full node, make transactions and audit the blockchain. Decisions on a federated blockchain need to be approved of the majority of members before being applied.

Federated blockchains are helpful, as they provide transaction privacy and are faster providing a much higher scalability than public blockchains without having a single point of failure, as private blockchains have. Pilkington (2015) sees them as a hybrid of public and private blockchains: “Partially decentralized, also called “consortium blockchains”, constitute a hybrid between the low-trust (i.e. public blockchains) and the single highly-trusted entity model (i.e. private blockchains)” . These blockchains are considered to be “partially decentralized” (Buterin 2015)

3.1.3. Private Blockchains

Private blockchains are managed by a single highly-trusted entity (e.g. company). Write and read permissions are controlled by the owner of the private blockchain. The owner is also the one validating every transaction within the private blockchain.

Private blockchains let the middleman/intermediary back in, as the owner of the private blockchain has full control of it and can even audit it, to reverse transactions etc. A likely application would be database management internal to a single company, where public readability might not be necessary (Buterin 2015). They are useful for use cases such as land registry, where the government would not recognize a land registry that it cannot control.

An advantage of private and consortium blockchains is that transactions become much cheaper, as they only need to be verified by a few high­processing nodes and they will be much faster processed . As read restrictions can be restricted a greater level of privacy can be provided.

3.2. Permissioned vs. Unpermissioned Blockchains

An unpermissioned blockchain does not have restrictions on identities of transaction processors whereas there is a predefined list of users with known identities who can perform transactions on a permissioned blockchain. A permissioned blockchain does not necesarrily need to be private as there are multiple levels of access to a blockchain:

"1. reading transactions from the blockchain, perhaps with further
restrictions (e.g., a user may have access only to transactions that involve him directly)
2. proposing new transactions for the inclusion into the blockchain
3. creating new blocks of transactions and adding them into the blockchain." (bitfury, public vs private ptl)

3.3. Consensus Mechanisms

The blockchain needs consensus amongst nodes, so the distributed ledger can be updated and the newest block can be added to the chain.

3.3.1 The Byzantine General Problem

Abbildung in dieser Leseprobe nicht enthalten

Fig. 4: The Byzantine Generals Problem. Only a coordinated attack of all generals at once can lead to success to conquer the city. The loyal generals (blue) want to attack the city, whilst the traitors (red) lied to the other generals about their plans and decide to retreat instead of joining the attack

The Byzantine General Problem is often used to illustrate the problems, that arise if one or more nodes on a Peer-to-Peer network are corrupt and might send conflicting messages or no messages at all. (Ruckeshauser 2017) How can the network still arrive at a good decision, if it is not known which nodes can be trusted and which nodes cannot be trusted? The Byzantine General problem is a made-up thought experiment, where multiple generals and their individual armies have surrounded a city to attack it. In order to have a successful conquest all armies need to attack the city at the same time. As the king has unexpectedly died the generals must decide by themselves how to act, as there is no authority that commands the generals. A multitude of problem arises for the generals, which all derive from low trust amongst the generals:

- the generals are several miles apart, they cannot see each other and communicate directly with each other in any way, so they need to use messengers to agree on a battle plan
- One or more generals could be traitors and send false messages to the others (or no message at all)
- one or more messengers could be traitors and send false messages to the generals
- the enemy city could intercept a messenger and forward a wrong message

So the generals must have a battle plan / algorithm that guarantees the following:

1. all loyal generals must decide on the same plan
2. a small number of traitors cannot cause the loyal generals to follow through a bad plan.

As in Peer-to-Peer networks such as blockchains the participants are in a similar low-trust situation as the Byzantine generals they need to have a distributed consensus that everybody agrees on in order to be functional. Blockchains have solved the Byzantine problem by using cryptocraphically secured consensus mechanisms.

Problems in Peer-to-Peer networks that are caused by one or more nodes behaving erratically are known as byzantine failures. As in such cases no consensus can be reached the network / the attack on the city is doomed to fail. Byzantine failures can be circumvented by the methods that are used by blockchains, namely: distributed ledgers and forgery-proof signatures.

In the Byzantine problem the application of a distributed ledger would mean that all generals repeat the messages they received to the others, thus making all messages publicly available. This way, a traitor who sent conflicting messages to different generals could be sorted out quickly. In (public) blockchains the records of all transactions are publicly available, which makes tampering with transaction data hard, as the intervention would be noticed by the other nodes. In blockchain this translates to the latest agreed-upon state / block, which is publicly readable and verifiable, as it is hashed so that editing it in any way afterwards would create a whole different hash that is invalid. In addition to that, this latest state is shared and distributed across multiple replicas in the network, so there is no single point of failure that could be attacked.

Forgery-proof signatures are another feature of blockchains, that can be used to solve the Byzantine problem. With those cryptographically secured signatures each participant holds a public and a private key. The public key acts as an address and is publicly available. The private key is secret and it is used to sign messages / transactions. The recipient of a message can verify if the message is actually from the apparent sender, making it impossible to tamper with the message after it is sent (Rückeshâuser 2017).

A (public) blockchain could not function without distributed consensus. It is a key feature of blockchains because it enables participants to interact with each other without knowing or trusting each other. In such a system transactions are secured by the consensus algorithm, instead of a highly- trusted third party or central authority. “Distributed consensus mechanisms are [...] not only responsible for the enforcement and execution of rules within the distributed ledger network, but are also responsible for enabling nodes to connect and work together, while tolerating partial failures” (Ruckeshauser 2017).

There are various distributed consensus mechanisms, that can be chosen depending on the all-encompassing goal of the distributed ledger / the blockchain. The decision also depends on the first layer (network layer). There are various trade-offs with each algorithm, as they all have differing performance and security and while some allow for some form of governance (especially in private blockchains) and others don't.

“A blockchain based system is as secure and robust as its consensus model” (Baliga, A. 2017). A failing consensus could lead to various unwanted outcomes:

- Blockchain Fork: A fork in the blockchain means that nodes in the network converge on different blocks as being part of the blockchain. This can also happen if two miners in Bitcoin find a nonce at about the same time. Such forks are temporary, as the protocol lets all participants eventually converge on a single chain. However, an unwanted blockchain fork can seriously negatively impact the network, as there is no consensus on a single state anymore and the view of data is inconsistent (Baliga, A. 2017)
- Consensus Failure: Some consensus mechanisms do not guarantee consensus under all conditions. E.g. if a majority vote is required and enough nodes fail or are acting maliciously a distributed consensus cannot be reached
- Dominance: Consensus is compromised if the outcome can be controlled by a single actor (or a group). This could be the case if the network is not resistant against so-called Sybil attacks, where the attacker creates many identities that are under his control. This
dominance could allow for confirming transactions that e.g. include double-spending transactions (spending the same coin twice). Dominance is also reached in Proof-of-Work blockchain if at least 51%of the miners decide to take over the network
- Poor performance: it might require too much time under certain conditions to reach a distributed consensus. This can happen if various nodes have turned malicious

The archival of consensus in a distributed low-trust system is difficult. “Consensus algorithms have to be resilient to failures of nodes, partitioning of the network, message delays, messages reaching out-of­order and corrupted messages. They also have to deal with selfish and deliberately malicious nodes” (WP-understanding).

Baliga (2017) identifies three key properties of consensus mechanisms for determination of applicability and efficacy:

1. Safety: A consensus protocol is determined to be safe if all nodes produce the same output and the outputs produced by the nodes are valid according to the rules of the protocol. This is also referred to as consistency of the shared state.
2. Liveness: A consensus protocol guarantees liveness if all non­faulty nodes participating in consensus eventually produce a value.
3. Fault Tolerance: A consensus protocol provides fault tolerance if it can recover from failure of a node participating in consensus.”

3.3.2. Proof-of-Work

The concept of the Proof-of-Work has been introduced in 1993 in order to fight spam mail, where the sender has to do some work that can be verified easily by the email provider (Schlatt, V. et al. 2016). Bitcoin uses the Proof-of-Work algorithm in order to validate transaction blocks and add them to the chain. It is the reason why the Bitcoin network needs so much electricity in order to operate. It is the most used consensus mechanism (also used by Ethereum, Litecoin and others) and relies on the so-called miners and financial incentives. In this scheme the miners need to proof, that they did a certain amount of work in order to qualify for the block reward. The miners compete in solving a complex mathematical puzzle that can only be solved by brute force, which means trying different combinations until the solution is found. In Bitcoin a number called 'nonce' has to be found. The node proposing a new block that gets added to the blockchain has to hash the proposed block in a way that the hash starts with a certain number of zeros (Ruckeshauser).

A SHA-256 one-way hash function is used. This means, that a message can be converted into a hash-value that has a length of 64 characters (random letters from A to F and numbers) looks somewhat like this SHA256(message) = ab530a1914982b79f9b7e3fba9... , but the message cannot be recreated through the hash value. In Bitcoin it is currently required to find a nonce that lets the hash start with 18 zeros, so that SHA256(message+nonce) = 000000000000000000b75e9164... . Chances to find the solution first increase with an increasing amount of computing power (and hence electricity) which has led to the existence of whole mining-farms - air-conditioned ware houses fitted with many computers and a sole purpose: mining crypto-currencies. The Bitcoin network performs currently ca. 30 quintillion (30*10¹⁸ ) hash calculations in each second (https://blockchain.info/de/charts/hash-rate?timespan=all). The number of calculations per second has increased exponentially in the last two years. At the beginning of 2016, the hash rate was at 0.8 quintillion calculations per second, whereas in 2014 it was at “only” 20 trillion (20*10¹² ). This really showcases the rapidly growing popularity of Bitcoin. The mathematical puzzle is asymmetric, which means that it is really difficult to solve and find the nonce but it is easy to proof. So that once a miner finds the right nonce, he announces it to the whole network. Once his solution is approved the blockchain is updated with the new block and the miner gets his blocke reward (on the Bitcoin blockchain it is currently 12,5 BTC). The mining serves two purposes: verifying new transactions and creating new tokens. As the incentives of the miners (making money by creating new crypto-currency) align with the interests of the blockchain network (maintaining the good operation of the overall system) the blockchain is able to work without a trusted third-party.

A blockchain with the Proof-of-Work consensus mechanism is under risk of a 51% attack. If more than half of the mining nodes are corrupt or rather if one entity has at least 51% of the networks hashing power, the network could potentially be taken over. Two cryptocurrencies already suffered such an attack (Varshney 2018).

With time the difficulty of the mathematical puzzle increases in order to align with the advances in technology and the growth of the network which lead to increasing computing power. As computing power is not a scarce resource per se, monopolizing is difficult which is for the best of the network.(Ruckeshauser 2017)

The resource-intensive process of mining in proof-of-work - secured blockchains is coupled with the obvious drawback of being wasteful on electricity. Specialized hardware is necessary for effectively mining Bitcoin, ASICs (Application specific integrated circuits) are designed with the sole purpose of mining. (Martindale, J. 2018) This effectively excludes the casual user from the mining and validation process, as it is not possible to do so with an ordinary personal computer or smartphone, leaving the mining to users and groups willing to invest in such hardware. It has led to a formation of mining-pools (multiple users cooperate to solve a block, the block reward is then split between them) and a concentration of mining pools in China, as electricity prices are really low there (blockchain.info).

Also, there are scalability problems with proof-of-work regarding throughput and latency (Vukolic 2015). The throughput is limited by the block size. Bitcoin allows a maximum of 6 transaction with its block size of 1MB. (Croman et. al 2016) This is really little compared to roughly 2000 transaction per second of Visa credit cards. (Visa.com 2016) which renders Bitcoin as it is now useless for replacing conventional payment systems in a bigger scale. The transaction time is bottlenecked by the inter-block time. In Bitcoin it takes 10 minutes to mine a block. A transaction usually has to be confirmed 6 times by the following blocks to be considered valid. (Croman et. al 2016). This means that it takes at least 60 minutes for a transaction to be finalized, which is also a concern when wanting to use Bitcoin as a regular payment system. The latency and throughput cannot be easily increased, as it would impact the security of the network.

However there are attempts to speed up the Bitcoin network and other PoW-secured cryptocurrencies in order to make it more scalable. One proposed update has been the “Segregated Witness”-update, which allows for more transactions to be added per block. It would make the implementation of the “Lightning”-network possible, which makes off- chain transactions possible. For this, a payment channel between two trusting parties will be opened, where instant transactions can take place. When the payment channel is closed, the new balances get added to the main-chain.(Croman et. al 2016). Litecoin already implemented the Segregated Witness-update, Bitcoin still has not done it (Preuss, M. (2017).

3.3.3. Proof - Of - Stake

Proof-of-Stake algorithms are designed to overcome the shortcomings of the Proof-of-Work algorithms. Namely, the high electricity consumption and the scalability issues of PoW were addressed with the PoS protocol. The whole mining operation of PoW is discarded and the chances of validating a new block and receiving the associated reward (which in PoS systems usually consists of transaction fees of all the transactions that happened within the block) depend on the node's ownership stake in the system (Pow vs PoS, bitfury). This is why it is sometimes called Proof-of- Ownership. The validator gets chosen randomly, chances increase with a bigger stake because a weighted random selection is used. “An individual stakeholder who has p fraction of the total number of coins in circulation creates a new block with p probability.” (bitfury, PoS). Also instead of just using the number of coins / stake in order to determine the next minter, coin age could be used, which rewards participants that have been on the network for a longer time. Coin age is the currency amount multiplied with the holding period (Kind and Nadal, Rückeshâuser S.173) “The rationale behind proof of stake is the following: users with the highest stakes in the system have the most interest to maintain a secure network, as they will suffer the most if the reputation and price of the cryptocurrency would diminish because of the attacks. To mount a successful attack, an outside attacker would need to acquire most of the currency, which would be prohibitively expensive for a popular system.” (BitFury Group 2015). This is another advantage of Proof-of-Stake compared to Proof-of-Work, as it is expected to be more expensive to gain the majority of the network's stake than to gain the majority of the network's computing power. Also, this malicious behavior could be detected by other participants and the funds of the attackers could be removed (Buterin 2015).

Furthermore, transactions are processed faster with PoS, and validating can be performed on smaller and weaker devices such as smartphones, as there is no need for much computational power and clients do not need to download the whole blockchain (Croman, K. et al. (2016). However, PoS systems are vulnerable to the so called “nothing-at-stake” - issue. Because there is no cost involved in mining at several chains (in PoW this is not the case, as mining twice as many chains would result in having to use twice the computer power) a node could vote for multiple blockchains and hinder reaching a distributed consensus. In such case double-spending would be made possible. In Proof-of-Stake secured blockchains, malicious behavior is punished. That could be the case if someone tries to revert a transaction. “The “one-sentence philosophy” of proof of stake is thus not “security comes from burning energy” (which is the case for PoW), but rather “security comes from putting up economic value-at- loss” (Buterin 2015). Instead of Proof-of-Work, where the miners secure the network due to financial incentives (block reward), under Proof-of- Stake the network is mainly secured by those who own the majority of coins, as they do not want to loose these it is under their best interest to maintain the network.

3.3.4. Delegated Proof-of-Stake

Delegated Proof-of-Stake (DPoS) is a system that is used by Bitshares and Tendermint. In this system blocks are minted (the equivalent of mining in PoW) by a predetermined set of users of the system (pos vs pow, bitfury). In some cases they get elected by the community and become representatives for their voters. In those cases the voting power of the individual users depends on their stake in the system. The delegates are rewarded for work (by collecting transaction fees) and punished for malicious behavior. In order to be a delegate they have to put funds into a “security deposit”. In case of bad behavior, those funds confiscated. Also, the delegates minting power (validating blocks) can be proportional to their security deposit. Those systems are referred to as deposit-based proof of stake (BitFury Group 2015) The delegates engage in two processes. One of them is building a block of transaction and the other is digitally signing a generated block in order to verify it. Blocks need to be signed by multiple delegates in order to be considered valid.

3.3.5. Proof-of-Authority

The Proof-of-Authority mechanism fits well for private and consortium blockchains. In this case certain nodes are selected as block validators when setting up the network. They will be trustworthy and usually have to provide some proof of identity. These authorities are in control of the network and are usually selected when setting up the network. “The main idea behind the PoA scheme is that blocks can be verified only by trusted signers”. (Tasca, P. et al. 2017). The “authorities” do not have to put capital at risk, like in Proof-of-Stake systems or stake economic resources (electricity) like in Proof-of-Work systems. Rather, their identity is at stake (POA Network 2017). As people have only one identity it is a scarce source and once their reputation is lost because of misbehavior it is much harder to regain it, than it is with coins in Proof- of-Stake secured systems. The PoA network (poa.network) uses this consensus mechanism. The “master of the ceremony” creates 12 initial key and distributed them to individual validators using a decentralized application. Those 12 validators can vote for removing or adding other validators using another decentralized application.

3.3.6. Proof-of-Burn

Proof-of-Bum requires the “burning” of digital coins in order to get mining rights for another cryptocurrency. Usually coins from a cryptocurrency that is secured by proof-of-work are sent to a random, veritably unspendable address, where they cannot be retrieved from, and thus are rendered useless. Compared to PoW, coins are burnt instead of electricity, whereas the idea is to give the newly mined coins the same value, that the burnt coins had, as electricity was spent on the burnt coins. (P4Titan 2014). By burning coins, a node gets “the permission to mine for a lifetime, which is actually a lottery between owners of burnt coins” (Ruckeshauser 2017). Slimecoin uses the Proof-of-Burn consensus mechanism and states, that the reason for using this form of consensus is removing the need for specialized hardware, allowing more people to participate in the network, as it is more feasible to “burn” coins compared to mining blocks in PoW. (P4Titan 2014)

3.3.7. Proof-of-lmportance

Proof-of-Importance is a blockchain consensus mechanism introduced by the cryptocurrency NEM. The consensus works similar to Proof-of-Stake, but includes more variables than just the amount of coins a node holds. The net transfers, meaning the amount if cryptocurrency spent in the last 30 days are also taken into the calculation of the “importance” of the individual note. Cluster nodes are valued more (nodes in interlinked clusters of activity) than outliers. Besides, a participant needs to have vested coins beyond a certain threshold to be eligible for adding a new block to the blockchain. (NEM Technical Reference 2018). Vesting means leaving coins in your account for a prolonged time. After every 24 hours 10% of “unvested” coins will become “vested”. The “vesting” is necessary to protect the network against Sybil attacks, where an attacker would create several fake identities and then transfer coins between those in order to increase his importance. But simply hoarding coins will not be very effective, thus the algorithm favors merchants and others that are active on the network and engage in many transactions. The benefit is, that those who use the network the most also profit the most while keeping the network secure.

3.3.8. Proof-of-Elapsed-Time

Proof-of-Elapsed-Time was developed by Intel for their blockchain project “SatwoothLake” which has become part of the open-source project HyperLedger of the Linux foundation. It mainly focuses on efficiency and requires a lot of trust towards the implemented TEE (trusted execution environment), which is the “scarce resource” in this scheme as opposed to electricity in PoW.

Each node in the network is assigned a random wait time and falls asleep for this time. The node who wakes up first is considered the “leader” is able to commit a block and notifies the other nodes. Assuming the waiting times have been produced truly randomly the PoET mechanism offers a fair lottery algorithm for all participants. As the waiting times are assigned by the nodes own CPUs using an enclave (a trusted function) cheating has to be prevented. Intel does so by interfering on a low level with the aforementioned TEE that runs on the node's processor. It is isolated against any interference from the operating system and thus cannot be manipulated from the outside (Sawtooth 2017).

3.3.9. Proof-of-Activity

Proof-of-Activity combines proof-of-work and proof-of-stake consensus mechanisms. It is designed in a way, that mitigates the risks of PoW and PoS systems, e.g. a 51%-attack (Bentov, I. et al. 2014). It extends the Bitcoin proof-of-work mechanism in a way, that more complex verifications have to be made. First, the miner tries to create an empty block header hash, which means that it only includes the miner's public address, the index of the pending block on the blockchain and a nonce. It does not include any transactions. If the hash of this empty block header is smaller than the current difficulty target it is broadcasted to the network. This hash derives N pseudorandom stakeholders, by concatenating the block’s hash with the hash of the previous block and with N fixed suffix values. Those combinations are hashed again and “follow the satoshi” is invoked, using those A hashes as an input. Satoshis are a one hundred millionth of a single Bitcoin, the smallest unit a Bitcoin splits into. “Follow the satoshi” means tracking these pseudorandom N satoshis down to their current owner, who is eligible for signing the block. All satoshis in existence up to the current block could be drawn. If a node owns twice as much cryptocurrency its chances are twice as big to be one of the lucky owners of the N satoshis. This is the proof-of-stake - like aspect of the proof-of-activity consensus mechanism because a bigger stake means a higher chance of signing a block. Assuming that all N drawn stakeholders are online and active, the last who signs it wraps the block up including all signatures and transactions for the hash of the entire block and broadcasting it to the entire network. Once it is proven to be a valid extension of the blockchain the transaction fees are shared between the miner and the A lucky stakeholders (Bentov, I. et al. 2014).

3.3.10 Proof-of-Capacity

With Proof-of-Capacity mining is only possible when allocating a sufficient amount of hard drive space, which is the “scarce resource” in this scheme (Tasca, P., Tessone, C. 2018). Those who dedicate more disc space have a proportionally higher chance of mining the next block and getting the block reward. This mechanism has aswell the benefit of being more energy-efficient than PoW and Burst—Coin, an adopter of Proof-of- Capacity also claims that it is fairer than Proof-of-Stake consensus because in PoS the richest are better than remunerated than the poor as they are able to stake more coins. As hard disks are available at low cost, anyone can easily become involved in the mining process (barriers to entry are lower than with other systems) (burst-coin.org 2018). PoC is similar in a way to PoW because in PoC the right nonce must be found. The difference is, that the miner is not performing a brute-force attack, trying as many combinations as possible, instead the miner pre-generates solutions using a very slow Shabal hash once, that are stored on the hard drive. This process is called “plotting”. The more disc space is allocated, the more nonces can be saved. The mining software reads through the cache for a few seconds with each block, scanning for the right solution. The rest of the time, the hard drive disc is in idle, waiting for the next block. (Tasca, P. et al. 2017).

3.3.11. Federated Byzantine Agreement

In this consensus mechanism each participant extends trust to a limited number of people within the network, which induces “flexible trust”, meaning that participants can choose freely whom to trust in order to reach consensus (Rückeshâuser 2017). This mechanism has been pioneered by the Ripple blockchain and is also used by the Hyperledger consortium. Before a transaction becomes added to the blockchain the participant has to wait until his trusted note confirm the transaction, and before that they have to wait for confirmation and so on. This whole group of nodes is called a quorum. In the Federated Byzantine Agreement, the consensus of a whole “quorum slice” needs to be achieved in order to add a block to the blockchain. Those slices can be set up according to individual rules e.g. it can be set up in a way, that the consensus of numerous quorum slices needs to be reached. Disjoint quorums (quorums, that do not intersect with other quorums) are undesirable for the network as they can act independently by simultaneously agreeing on contradictory transaction which undermines the overall consensus. (Mazières, D. 2017)-

The advantages of the federated byzantine agreement are that anyone can join with a low barrier to entry the system, as the membership is open and control over the network is decentralized, while the nodes can individually choose who to trust in the network, making it possible to trust multiple quorum slices or even basing choices on external criteria (e.g. a landlord could require acknowledgement from a trusted bankind node, a trusted credit agency and a node associated with the department of housing).

3.3.12. Practical Byzantine Fault Tolerance

This is one of the first solutions to the byzantine generals problem and was developed in 1999 (Castro, M., Liskov, B. 1999). It is currently used by Hyperledger Fabric with networks that consist of less than 20 participants. The consensus mechanism tolerates up to one third of participants behaving erratically. That means, they could send misleading messages, not respond at all or really try to undermine the whole system. It does not matter with PBFT, as long as two third of notes are still honest an overall consensus can be achieved.

Breaking the consensus it down it basically works like this: if a message/transaction is sent, every node that receives the message cryptographically signs it and then broadcasts the received message again to the whole network. This way, conflicting messages can be identified and the traitor can be found (Kwon, J. 2014). Tendermint uses a modified version of PBFT, which allows for more participants and is really efficient In this scheme two thirds of consensus are needed in order to approve a proposal. Tendermint is very fast and accepts any programming language, it is really promising as this consensus can combine consensus over multiple different blockchains, making “living in a multiple blockchain world” more feasible (Swan 2017)

3.4. Smart contracts

Smart contracts where first proposed by Nick Szabo back in 1996 (Szabo 1996). Szabo suggests that the contract is the basic building block of a free market economy. Societal norms such as property rights and the structures that go along with it have developed over several centuries and are worthwhile keeping. In order to handle those contracts in the cyberspace era he came up with smart contracts. They are called “smart” because they are way more functional than their paper-based ancestors. Ethereum has been invented to expand the possibilities of crypto-currency beyond just being a value holder With smart contracts, digital assets can be moved automatically according to pre-specified rules (Buterin 2013). Basically, a smart contract is an agreement between parties without the need for a third party because smart contracts execute themselves if the pre-specified rules are met. For example payments could be automated with regards for several constraints e.g. “A can withdraw up to X currency units per day, B can withdraw up to Y per day, A and B together can withdraw anything, and A can shut off B's ability to withdraw". (Buterin 2013). Neither party is able to violate the terms of agreement specified in a smart contract and because they run on the blockchain they are tamperproof. Put in a simple form, smart contracts rely on the IF ^ THEN principle, meaning than once the IF-condition is met the THEN- condition is executed. The list of possible use cases for smart contracts is very extensive, ranging from supply management (if a supply runs low, new supply is ordered) over automatically executed payments to managing smart properties (e.g. lending of an apartment or vehicle can be automatically and autonomously managed by smart contracts).

3.5. Decentralized Applications

Decentralized applications (DApps) are run on top of a blockchain instead of a single server or system. This protects them againsThey use smart contracts in order to access and communicate with their underlying blockchain. They have both, front and back end code, offering a user interface. DApps are open-source, operate autonomously and do not have any controlling entity. Changes within the DApps require consensus amongst stakeholders. All the data is cryptographically stored in a publicly accessible blockchain. DApps usually have tokens that allow access to the network and contribute value to the participants. Mining is enabled through a cryptographic algorithm like Proof-of-Work. There are three types of dApps: Type 1 DApps have their own blockchain (e.g. Bitcoin, Ethereum) , whereas a Type 2 DApp is a protocol relies on the blockchain of type 1 Dapp (e.g. most smart-contract based apps that use Ethereum). Then again a type 3 Dapp uses the protocol of a Type 2 DApp. (Johnston, D. et al. 2016). This all sounds really confusing, but drawing an analogy to “regular” applications helps. In this case the type 1 application would be the operating system, like Windows 10. A type 2 application for the operating system would be the web-browser or a text­processing program. An add-on (e.g. add-blocker for the web-browser) would be the type 3 application.

Ethereum founder Vitalik Buterin separates DApps into three main categories (Ethereum white paper):

- Financial applications: Providing users with methods to manage finances (fiat and crypto-based) including financial derivatives, saving wallets and even full-scale employment contracts
- Semi-financial applications: Mix money with information from outside the blockchain (e.g. insurance apps, that refund money for plane tickets if flight is delayed)
- Non-financial applications: For example identity verification, online voting, decentralized governance or decentralized file storage system.

4. Blockchain Use Cases

There is a remarkably broad spectrum of blockchain use cases. It can virtually be applied anywhere, where humans interact with each other. The most obvious use of blockchains is the use of it as a value holder or rather as a currency. The financial use of blockchains is also the most researched use case for them, as the vast majority of blockchain applications are cryptocurrencies. However, there are really interesting non-financial uses of blockchain technology, that can change the ways we interact with each other. It has a great potential for dismantling bureaucracy and making information exchange more efficient, secure and transparent for all involved parties. A few select use cases are going to be briefly discussed in the coming chapters.

4.1. Currency

With Satoshi Nakamoto's brain child Bitcoin the first ever blockchain was publicly deployed. His idea found many supporters and people who invented their own version of a cryptocurrency. It is an excellent idea, especially in public and permissionless blockchains, so that anybody with a smartphone and an internet connection canjoin. This way people who do not have access to banks can have access to money and possibly credit through cryptocurrency (Kuznetsov 2017).

But can conventional fiat currencies really be replaced by cryptocurrencies? Cryptocurrencies do not have any underlying value but fiat currencies are also just a piece of paper whereas digital coins are just chunks of data. Transactions can be processed much faster and with lesser fees, as there is no intermediary between the sender and the receiver in blockchain-based currencies. So far, there is still quite limited real-world use, as only a few businesses accept Bitcoin as a form of payment. Furthermore, cryptocurrencies are subject of huge volatality which hinders a more general acceptance. The negative environmental impact of proof-of-work based currencies is also an issue, that could be remedied by using other consensus mechanisms or switching over to an alternative like GridCoin, which steers the computational power towards scientifically useful work (bravenewcoin.com 2017) However, blockchain and its applications are still in their early stages, only time will tell if cryptocurrencies become more stable and widely adopted.

4.2. Digital Identity

Blockchain-powered services like SecureKey, which is based on a permissioned Hyperledger blockchain, provide ways of verifying an individual's identity online. Blockchain has the “capacity to reduce cost and fraud and to simplify the experience for customers while also keeping out the bad actors” (Wolfond, G. 2017). As more users migrate to digital platforms, there is a need for trusted digital identities. A modern approach towards a trusted digital identity would greatly reduce friction and transaction costs. It would benefit all parties such as governments, banks,

insurances, healthcare providers by reducing administrative expenses and it would benefit all individuals who will be able to conveniently proof their identity and allowing them to have control over their own personal data (Wolfond, G. 2017).

4.3. Supply Chain Management

By supplying consumer goods with some form of digital identity blockchain-based supply chains could be introduced, providing a tamperproof history of product manufacturing, handling, maintenance and a registry of all stations a good goes through until it arrives at the costumer. It would remove the single point of failure because the data would be stored in a decentralized manner. Supply chain transparency would be ensured and fraud would be reduced. There are various start-ups working on blockchain solutions for supply chain management such as Everledger that deals with diamonds, allowing to track a diamonds history and thereby ensuring it was not stolen or a blood diamond, which finances conflicts in Africa (Lee and Pilkington 2017)

4.4. Digital Voting

By providing citizens with a trusted digital identity, digital voting could be established easily. Blockchain technology has the right properties for voting processes to reduce fraud and providing security, such as being tamperproof, immutable and inaudible. Nowadays voting is not very efficient, with every vote hand-counted. Blockchain-enabled online voting would cut voting costs greatly and possibly raise voter participation on a higher level because online voting will be far more convenient for citizens than taking the time to make it to their polling place.

4.5. Health Data Management

This use case would also go hand in hand with a digital identity. A blockchain within the health sector would guarantee frictionless sharing of patient-information between doctors, insurances, hospitals and pharmacies. There are already various start-ups that invented blockchain- solutions for the healthcare sector such as Scalamed and Medicalchain that have different approaches (public and private blockchains). By the use of smart contracts stakeholders are only able to access information to which they are specifically entitled to (Engelhardt 2017). The health data could possibly even go beyond sharing informations between healthcare providers. A “new age of healthcare” might be around the corner. BurstlQ aims to integrate data streams to gather new insights into individual best health outcomes and help realizing those goals. (Engelhardt 2017).

4.6. Sharing Economy

“Sharing” - services such as Airbnb could be run on top of a blockchain. Using a decentralized self-governed application that uses smart contracts for payments and allowing the lending of a certain objects. That would cut out the middle man and thereby lower costs “Smart objects” (e.g. a flat or car that will open itself once it is paid for) could be implemented and make the whole process autonomous. The German start up slock.it works on providing such self-renting objects.

5. Conclusion

In this thesis I showed how blockchain technology functions and identified the system design features it consists of. So how do they impact your use case? It really depends on many variables and even for the same use case different solutions can be applied (e.g. public and private blockchains for the same use case in health data management). The very first question when designing a blockchain should be whether a blockchain is really needed for a specific use case or not. Wust and Gervais (2017) managed to come up with this decision diagram:

Abbildung in dieser Leseprobe nicht enthalten

Fig. 5: Do you need a blockchain? (Wüst and Gervais 2017)

Going back to figure 2, that displays the four-layer architecture of blockchains, the first layer is the network layer and when building your own blockchain, choosing the type of network will most likely be the first decision to make. Blockchain networks can be either public or private and they can be permissioned or unpermissioned. Desired anonymity and trust in validators are the decision variables for establishing which type of network to build your blockchain on.

Figure 6 gives a general idea what type of network a blockchain should be based on depending on the desired anonymity and established trust in the validators. Permissionless public networks make mostly only sense for cryptocurrencies and not so much for most other use cases because establishing the individual's identity is important most of the times in order to make legal contracts about physical property (e.g. sharing economy) or to verify that the individual is no one else (e.g. voting). Scalability is also an issue with the big permissionless and public networks. One possible solution for this would be “sharding”, that Ethereum founder Vitalik Buterin recently teased on Twitter. In this scheme the blockchain is split into several shards, so that every validating node would only have to carry a part of the network as opposed to the whole network.

Abbildung in dieser Leseprobe nicht enthalten

Fig.6: What type of blockchain to choose depending on the desired anonymity of the users and the trust in the validators. Adapted and modified from Kravchenko (2017)

The second layer of architecture in blockchains consists of consensus algorithms. As I layed out in this thesis there are many consensus mechanisms to choose from. They all have various trade-offs and it might be difficult to choose one. Note, that hybrids are also possible e.g. of proof-of-stake and proof-of-work designs.

Abbildung in dieser Leseprobe nicht enthalten

Fig. 7: Comparison of consensus mechanisms (Tasca 2018)

Blockchain is still in a state of infancy with many different blockchain start-ups emerging. It might take decades for the technology to take over , as it is a rather foundational than disruptive technology. With services like Polkadot it is possible to connect different blockchains to each other, which makes it more realistic for a blockchain-powered cryptoeconomy to gain traction and find its way into the mainstream. An exciting future awaits the blockchain, only time will tell whether the technology is going to become an integral part of our economy.

References

Babbitt, D., Dietz, J. (2014). Crypto-economic Design: A Proposed Agent-Based Modelling Effort, University ofNotre Dame, USA.

Baliga, A. (2017). Understanding Blockchain Consensus Models. Pune, Persistent Systems.

Bentov, I. et al. (2014). Proof of Activity: Extending Bitcoin’s Proof of Work via Proof of Stake.

BitFury Group (2015). Proof of Stake versus Proof of Work.

Blockchain.info (2018): Hashrate Verteilung. URL: https://blockchain.info/de/pools. 28.4.2018.

Blockchain.info (2018): Hash Rate. URL: https://blockchain.info/de/charts/hash-rate?timespan=all. 15.3.2018.

Bouillon, M. (2017). Disruptives Potential der Blockchain-Technologien und die Möglichkeit dezentraler Zahlungssysteme, URL: https://www.grin.com/document/352159. 12.2.2018.

Buchman, E. (2016). Tendermint: Byzantine Fault Tolerance in the Age of Blockchains, Guelph, Ontario, Canada.

Burst (2018): Blockchain, URL: https://www.burst-coin.org/proof-of- capacity, 3.5.2018.

Buterin V. (2013). Ethereum White Paper-ANext Generation Smart Contract & Decentralized Application Platform.

Buterin, V. (2013): What Proof of Stake Is And Why It Matters. URL: https://bitcoinmagazine.com/articles/what-proof-of-stake-is-and-why-it- matters-1377531463/, 2.4.2018.

Buterin, V. (2015): On Public and Private Blockchains. URL: https://blog.ethereum.org/2015/08/07/on-public-and-private-blockchains/, 5.4.2018.

Buterin, V. (2017): Introduction to Cryptoeconomics, URL: https://vitalik■ca/íiles/intro_crvptoeconomics■pdf. 26.3.2018.

Castro, M., Liskov, B. (1999): Practical Byzantine Fault Tolerance. In: Proceedings of the Third Symposium on Operating Systems Design and Implementation, New Orleans, USA.

Croman, K. et al. (2016). On Scaling Decentralized Blockchains.

Financial Cryptography and Data Security, pp. 106-125.

Garzik, J., Bitfury (2008): Public versus Private Blockchains, URL: http://bitfury.com/content/downloads/public-vs-private-ptl- l.pdf. 7.4.2018.

Gordon,R.J.(2017). TheRise and Fall of American Growth: The U.S.Standard of Living Since the Civil War.PrincetonUniversity Press.

Halford, R. (2014). Gridcoin - Crypto-Currency using Berkeley Open Infrastructure Network Computing Grid as a Proof Of Work.

Iansati, M., Lakhani, K.R. (2018). The Truth About Blockchain, Harvard Business Review.

Johnston, D. et al. (2016). The General Theory ofDecentralized Apps, Dapps.

Karen, H. (2018). Living in a multiple blockchain world. In: Henry Stewart Publications: Journal ofDigital Banking, Volume 2, Number 3, pp. 223-231.

Kravchenko, P. (2017): Consensus explained, URL: https://medium.com/@,pavelkravchenko/consensus-explained- 396fe8dac263, 17.4.2018.

Kuznetsov, N. (2017): How Emerging Markets And Blockchain Can Bring An End To Poverty, URL: https://www.forbes.com/sites/nikolaikuznetsov/2017/07/24/how- emerging-markets-and-blockchain-can-bring-an-end-to- poverty/#74e4dd204a0c■ 1.4.2018.

Kwon, J. (2014). Tendermint: Consensus without Mining.

Lee, J., Pilkington, M. (2017). How the Blockchain Revolution Will Reshape the Consumer Electronics Industry. In: IEEE Consumer Electronics Magazine, July 2017, pp. 19-23.

Martindale, J. (2018): What is an ASIC miner?. URL: https://www■digitaltrends■Com/computing/what-is-an-asic-miner/. 22.4.2018.

Mazieres, D. (2017). The Stellar Consensus Protocol: A Federated Model for Internet-level Consensus, Stellar Development Foundation.

McPhee, C., Ljutic, A. (2017). Editorial: Blockchain. In: Technology Innovation Management Review, Volume 7, Issue 10, pp. 3-5.

Nakamoto, S. (2008). Bitcoin: A Peer-to-Peer Electronic Cash System.

NEM Technical Reference (2018).

P4Titan (2014). Slimcoin A Peer-to-Peer Crypto-Currency with Proof-of-Burn.

Park, S. et al. (2015). SpaceMint: A Cryptocurrency Based on Proofs of Space, MIT.

Perfall, A. et al. (2016). Blockchain - Chance für Energieverbraucher?, Verbraucherzentrale NRW.

Pilkington, M. (2015). Blockchain Technology: Principles and Applications, Edward Elgar.

POA Network (2017): Proof of Authority: consensus model with Identity at Stake, URL: https://medium.com/poa-network/proof-of-authority- consensus-model-with-identitv-at-stake-d5bd15463256. 16.3.2018.

Preuss, M. (2017): Litecoin aktiviert erfolgreich SegWit - Undjetzt?. URL: https://www.btc-echo.de/litecoin-aktiviert-erfolgreich-segwit-und- jetzt/, 11.3.2018.

Rückeshäuser, N. (2017). Distributed Ledgers for the Prevention of Accounting Fraud, Freiburg.

Schlatt, V., Schweizer, A., Urbach, N., Fridgen G. (2016). Blockchain: Grundlagen, Anwendungen und Potenziale. Bayreuth, Fraunhofer FIT.

Schlatt, V. et al. (2016). Blockchain: Grundlagen, Anwendungen und Potenziale, Fraunhofer FIT.

Sawtooth (2017): PoET 1.0 Specification, URL: https://sawtooth.hyperledger.org/docs/core/releases/1.0/architecture/poet.h tml, 17.4.2018.

Swan, M. (2017). Anticipating the Economic Benefits ofBlockchain, Freiburg.

Szabo, N. (1996). Smart Contracts: Building Blocks for Digital Markets.

Tasca, P. et al. (2017). Ontology ofBlockchain Technologies. London.

Tasca, P., Tessone, C. (2018). Taxonomy ofBlockchain Technologies. Principles of Identification and Classification, London.

The Economist (2015): The next big thing. URL: https://www.economist.com/news/special-report/21650295-or-it-next-big- thing, 13.4.2018.

Unibright.io (2018). Blockchain Evolution: from 1.0 to 4.0, URL: medium.com/@UnibrightIO/blockchain-evolution-from- 1-0-to-4-0- 3fbdbccfc666, 1.5.2018.

Varshney, N. (2018): Proof-of-Work tech under fire after 51% attacks on Electroneum and Verge. URL: https://thenextweb.com/hardfork/2018/04/05/proof-of-work-tech- electroneum-verge. 8.4.2018.

Vukolic, M. (2015). The Quest for Scalable Blockchain Fabric: Proof-of- Work vs. BFT Replication. Zurich, IBM Research.

Wolfond, G. (2017). A Blockchain Ecosystem for Digital Identity: Improving Service Delivery in Canada's Public and Private Sectors. In: Technology Innovation Management Review, Volume 7, Issue 10, pp. 35­40.

Wüst, K., Gervais, A. (2017). Do you need a Blockchain?, ETH Zürich, Switzerland.

Visa.com (2016): How a Visa Transaction Works. URL: http://web.archive.org/web/20160121231718/http:/apps.usa.visa.com/mer chants/become-a-merchant/how-a-visa-transaction-works.jsp, 23.3.2018.

List of Figures:

Fig. 1: Traditional Transactionmodel compared to blockchain Transactionmodel. In blockchain networks transactions are stored and encrypted on the distributed ledger which is updated by consensus amongst participants. Smart contracts are individually defined rules (e.g. about prices and quantities) demand meet supplly autonomously.

Fig. 2: The four-layer architecture of distributed ledgers. Adapted and modified from: Nadine Ruckeshauser (2017)

Fig. 3: An exemplary blockchain and merkle tree graphically illustrated. Block 27 points to the previous block by having stored the matching hash value. The matching nonce for the merkle root has to be found by miners so that the block can get added to blockchain

Fig. 4: The Byzantine Generals Problem. Only a coordinated attack of all generals at once can lead to success to conquer the city. The loyal generals (blue) want to attack the city, whilst the traitors (red) lied to the other generals about their plans and decide to retreat instead of joining the attack

Fig. 5: Do you need a blockchain? (Wüst and Gervais 2017)

Fig.6: What type of blockchain to choose depending on the desired anonymity of the users and the trust in the validators. Adapted and modified from Kravchenko (2017)

Fig. 7: Comparison of consensus mechanisms (Tasca 2018)