How it works

At its core a blockchain is nothing more than a database shared between participants in a network without a central governing authority. How can you get everyone to agree on what its state should be? Imagine this database contains account balances, as in the case of cryptocurrencies. If everyone owns a digital copy of it where they are free to change whatever they want (it is stored on their own computers after all), what prevents them from manipulating it in their favour? What is needed is some kind of consensus protocol that allows everyone to agree on one version of this “database”. This is the fundamental idea behind the blockchain.

Reaching consensus in a decentralised network

In a blockchain, records, i.e. updates to the database, are batched into “blocks”. In addition to this new information, each block contains a timestamp and reference to the previous block, hence the name “blockchain”.

The integrity of each block is verifiable via its hash, a number that is the result of a computation on the entire block. Hashing algorithms have the property that any tiny change to their input produces a completely different output, so any change to a block would result in a different hash, making it impossible to manipulate a block once it is “buried” in the chain, since all successive blocks depend on the previous ones.

This solves only part of the problem, however. It only ensures that information in the blockchain is immutable once it is stored, meaning it cannot be changed without tampering evidence. In addition, a mechanism is required to ensure that only correct new information makes it into the blockchain, i.e. a way of democratically deciding what information should be added in the next block.

How should voting work in a digital system? This depends on what kind of blockchain we're talking about.

Different kinds of blockchain

While many different blockchain implementations have been developed, they can be divided into two types:

• Permissionless: In a permissionless blockchain, anyone can read the contents of the shared database and join the network and act as a validating node, i.e. participate in the voting process. The most well-known examples of permissionless blockchains are Bitcoin and Ethereum.
• Permissioned: access to the network (and thus voting) is restricted to a set of known participants. Ripple is one such blockchain.
If outsiders are allowed to read and verify the integrity of the ledger's contents, we speak of a 'public permissioned blockchain', if this is not the case, it is referred to as a 'private permissioned blockchain'.

In a permissioned blockchain, voting is relatively straightforward to implement since it is precisely known who is part of the network. When this is not known, as in the case of a permissionless blockchain, things get tricky: A plain “one node - one vote” principle would be absolutely useless, since there is no authority keeping track of participants and consequently, nothing to prevent anyone from joining multiple times to increase their voting share. The first solution to this problem was proposed by the creators of Bitcoin: Instead of “one node - one vote”, the Bitcoin consensus protocol employs a “one computation - one vote” mechanism. What this means is that voting power is determined by the computational resources a participant is devoting to the network in a so-called a “proof-of-work” (PoW) algorithm. So long as these resources remain distributed with no single entity controlling more than 50%, the algorithm ensures that nobody has the power to manipulate the system in their favour.

One major drawback of such algorithms, however, is that they only fulfil their task by being inherently wasteful. According to one estimate, the Bitcoin network uses as much electricity as the all of Israel.[1] Another issue arises when the specific PoW algorithm used enables an economy of scale, meaning that network participants can increase their voting power more than linearly by dedicating more resources. This leads to the network naturally tending towards centralisation.[2]

A newer approach is to make use of a “proof of stake” (PoS) protocol. Voting power in such a system is determined by a participant’s “stake” (ownership of digital assets, i.e. the cryptocurrency issued on the network), rather than physical assets (computers and electricity to perform the computations required by PoW). As such, there is a disincentive for acting against the common interest, since doing so would diminish confidence in the system and cause the value of the digital assets to drop, thereby hurting the most the very ones with the most voting power. It is also less susceptible to the problem of economies of scale outlined above. However, it is problematic in a different way: If it is easy to compute blocks, it opens up the possibility for malicious participants of precomputing alternative versions of the blockchain. When new nodes, or those that have been offline for some time, have to decide which version of the blockchain they accept as valid, it is not as straightforward to do so as in a PoW scheme, where it is simply the longest chain (the one that overall had the most resources dedicated to it).

Whichever way consensus on the next block is reached, a key property of blockchain technology is that it is easy for other participants in the network to check that a new block conforms to the protocol before they add it to their own copy of the chain and the cycle restarts.

Cryptocurrencies

Now that we have an overview of how the blockchain works in general, we can look at its original application in cryptocurrencies.

The promises of cryptocurrencies

Most of the excitement around cryptocurrencies stems from their perceived potential of providing a better alternative to normal currencies and centralised online payment systems in several ways, but as we will see only some of them actually hold up under scrutiny.

The first of them is freedom to spend. Since no central authority has control over them, cryptocurrencies based on permissionless blockchains can’t restrict who can join and who can make transactions to whom. As such, they are the digital equivalent of cash. This resonates with the common libertarian spirit in the crypto-community that movement of capital should be completely free from third-party interference (e.g. legal regulations).

Another one is anonymity (or at least pseudonymity in the case of most cryptocurrencies). Since “opening” a cryptocurrency account involves nothing more that the click of a button to generate a new key pair (versus the KYC “know your customer” regulations banks have to adhere to), it enables users to hide their identity behind multiple identities on the blockchain. However, since the blockchain itself is a completely open book holding the entire record of all transactions ever made on it, it is up to the users to make sure they don’t reuse addresses if they want to stay anonymous.

Safety from inflation is yet another advantage pointed out by proponents of cryptocurrencies. Since their supply rate is predetermined and their total supply ultimately capped, cryptocurrencies cannot “fall victim” to monetary policies that increase the supply of money such as quantitative easing. Again this resonates well with libertarian world views.

Last but not least, cryptos promised to have lower transaction fees than centralised payment system providers such as credit card companies and PayPal. The logic behind this argument was that since no company was behind them needing to take a cut as profit, the newly released coins generated by solving blocks were enough incentive for miners to keep the system fee-free for users. However, this has not held entirely true. While fees are not set by a central authority, miners prefer transactions that have a higher fee attached to them. In this way, users have to pay a higher fee if they want their transaction to be processed faster. Also, since block size is limited, the transaction fee depends not on the amount of the transaction but its size in terms of kilobytes. So while PayPal charges up to 4.4% of the transaction amount plus a fixed fee of ca $0.50 [3], the cost of a Bitcoin transaction depends on the network load. As such, it has peaked at more than 35$ in December 2017 and now varies between $0.20 and $0.60.

Do you really want to pay in crypto?

While properties such as anonymity and the freedom to evade legal restrictions on capital flows are attractive to some, cryptocurrencies come with their own limitations. For example, the core property of the blockchain being immutable, making transactions final, might turn out to be not as attractive to the average law-abiding user after all. If they fall for a scam, there is nobody to call to reverse the transaction (which they could do had they used a credit card). If they lose their private keys or someone hacks their wallet, the associated coins are gone forever thanks to the power of cryptography.

The scalability issue

This issue is at the heart of current cryptocurrency development. It refers to the limited throughput capacity of such decentralised systems. Currently, Bitcoin is able to process up to 7 transactions per second (tps).[4] Ethereum can handle around 15tps, which might seem like a considerable improvement but still pales in comparison to the capacity of centralised payment providers such as Visa, which can process up to 45,000tps.[5] How can it be scaled up to match this?

The fundamental limiting factor here is the block size. In the case of Bitcoin, it is artificially set to 1MB. However, increasing it by a factor of 1,000 is not an option, since a block size of 1GB would be impossible to handle for most nodes. [6]

For this reason current developments to overcome this issue focus on so-called off-chain solutions, such as the Lightning network for Bitcoin and Raiden for Ethereum. The idea is that instead of recording every transaction on the blockchain, participants can set up direct payment channels between themselves. Essentially a payment channel is a smart contract either one or both participants deposit funds in. It works as follows: Imagine you want to pay for your daily coffee in Bitcoin/Ethereum. It would be quite inefficient to record a transaction of 2 pound equivalent on the main blockchain. Instead, you and your coffee shop would simply keep a record of when and how much funds have moved between you in your channel, signing it every time. Since the rest of the network wouldn’t need to be aware of these transactions, they could happen near instantly and for free.[7] Transaction fees would only need to be paid when the channel is opened and closed, which is when the final balances are recorded on the main chain. But what if your friend visiting from abroad wants to pay for your coffee next time? If they don’t have a pre-existing channel with the coffee shop, opening a new channel just for one transaction wouldn’t solve the original problem, since the opening of a channel is equivalent to a regular on-chain transaction. Here’s where it gets really interesting: If you have a channel with both your friend and the coffee shop, the network will route the transaction via you in the middle. This idea could then be scaled up to route payments via several intermediaries, drastically reducing the load on the main chain.

However, there are some problems with this solution. First of all, there is some risk of fraud. If a participant of a payment channel decides to close the channel, they can submit an older version of the transaction record in which their side of the balance sheet was higher. The other party would then need to make sure to submit the newer record within a certain period of time to ensure none of their funds are lost. [8]

Another issue is that since everyone is essentially lending funds to each other, the capacity of a path from A to B is limited by the channel with the smallest amount of funds locked in it. This leads to a certain conundrum: The more payment channels exist, the smaller the capacity of each one will be, the fewer there are, the more centralised the system becomes, with larger entities having many channels serving as liquidity providers. [9]

Same same but different

So far we have restricted our discussion to the current two main players in the field, namely Bitcoin and Ethereum. However, thousands of cryptocurrencies exist, collectively termed "alt coins". Before we move to the next section, let us take a moment to briefly look at two of them and what makes them different:

Monero

As mentioned before, Bitcoin and Ethereum are not anonymous, rather pseudonymous, meaning every transaction is openly recorded in the public ledger under the blockchain. Monero's main aim is to improve anonymity by obscuring both sending and receiving addresses, making it impossible to trace the flow of coins.

Cardano

Cardano is a third-generation cryptocurrency aiming to make further improvements to the innovations that Ethereum brought as a second-generation, namely native support for smart contracts. It is based on academically peer-reviewed open source code.

So, do you need a blockchain for it?

Cryptocurrencies, being the original use case of blockchain, obviously depend on it, so the answer is 'yes'. However, the viability of cryptocurrencies themselves is a more difficult question, as outlined above.

IoT and Blockchain

IoT can be described as a network between physical objects which are connected to the internet and generate a great deal of data using different types of sensors. With the rise of all sorts of objects being connected to the internet such as phones but also fridges, the IoT grows. Communications increase. There are more and more concerns about the security of IoT solutions.

Can we combine IoT and Blockchain?

Blockchain adding value to IoT

Firstly, blockchain can be very useful in various IoT applications and add value to them. What is important to understand in an IoT application is that the data has its own life cycle. It is firstly created, travels between different objects, it is exploited and ends up in a finite state. The different objects taking part and the transactions of data must be verified. Furthermore, the data that is shared must be secure. Blockchain could be very useful since it would record permanently transactions between different objects, their states and the data they are generating.

We will now take the example of a smart home in order to illustrate this idea. Objects within the house need to communicate between each other by passing on data. Much of this data is personal because it has details about our lives and daily routines. But this data moves and hence there are far more openings for hackers to potentially attack us. A solution could be to construct a permissioned blockchain in which the only participants of this blockchain would be the objects in the smart home. This reduces a lot of risks. A blockchain like this means that no malicious object can communicate with the smart home and sent erroneous data. In addition, sharing of the data is made secure and personal data stays within the blockchain. Owners and participants could send instructions as a transaction. [11] This solution would be feasible but quite expensive since the environment of a smart home is quite small. Comcast Labs are currently building a new model within the smart home area. They have conceived a permissioned-blockchain for customers to be in total control of their digital homes. With this solution, they could easily grant, revoke and taylor access to any IoT device to family members or guests. Only a limited set of people have access to the blockchain and to different objects. For example, parents may want their children to have access the garden but not to the thermostat. This could be achieved by sharing private keys of specific objects.[12]However, this project seems a bit ahead of its time and has not been build yet. We could also consider a blockchain for a smart city. Indeed, this IoT application is at a much larger scale than a house and the data that is being generated is much larger. There are more chances that security breaches take place and sharing of data between different objects or organisations is key to a smart city but also to any IoT solution. Therefore the blockchain would be even more valuable since many more objects can interact and the blockchain is distributed within all of them. The data pushed onto the blockchain could regard traffic congestion, pollution, electricity flow or water distribution for example. The blockchain is then a platform for the object to share live data from different locations. The boundaries of an idea like this is the time taken to verify the transactions. For traffic flow for example, we would need the data to be near-instantaneous for it to be useful for other drivers. IoT and blockchain are being combined and solutions are created : IBM has created a platform that lets you add selected IoT data to a private blockchain. The data is shared only among the business partners involved with the transactions.

IoT adding value to the blockchain

On the other hand, we can see that IoT may add value to a blockchain. The idea is to use IoT devices to build a secure and transparent and error-free blockchain. Indeed, connected objects offer real time data that can be exploited or shared. If we take the example of the food chain, then we may be interested in food safety. The solution would be to build a blockchain which would store the different variables that have requirements such as temperature, humidity or light for a specific product. This recording would take place across the chain from the raw material to the product, packaging and transport. However, the questions we could ask ourselves is how can we generate this data and how can we ensure that the information going onto the blockchain is accurate and trust worthy. We could believe humans but the use of IoT objects would allow more trust and less errors. The IoT devices would need to come from a tamper proof end point that is sending data on a periodic basis. IoT sensors can exchange data through the blockchain(one can argue that a certain party can take control of the IoT devices and the values they return. A blockchain may not be necessary if we need a third party to make sure the IoT devices are neutral).[13] Since they are no intermediation, the cost of IoT can be brought down. It adds a lot more trust and automation to the process. We would then have a record across the entire food chain of the conditions of the product. The data could either be verified at the end of the chain or during it.

Another idea combining blockchain and IoT related to the food chain would be the tracking of data regarding an object that has a long-life expectancy such as a car. It is always a gamble to buy a second-hand car. Hence, we could implement a blockchain between the manufacturer, the user and the car maintainer and any other actor which has an impact on the car’s state. The sensors implemented on the car could push data onto the blockchain such as average speed and hard-breaking. Furthermore, if the manufacturer decides to update the car’s software, then this piece of information could also be pushed. This blockchain would allow a record of the car’s history and could also be of use to the insurer. The main advantage of a blockchain is the distributed data of the car to the participants which are all entitled to it, and the secure storage of the history.

So do we need a blockchain?

The combining of blockchain and IoT, would ensure trust, safety but also a standardized solution to many of today’s challenges. Blockchain is useful for the sharing, storing, and adding of data. The IoT on the other hand, is convenient for data creation and representation of the state of some process. The combining of the two technologies has definitely potential. On the other hand, projects of today may be a bit ahead of their time. There are issues regarding the people controlling the connected objects, the time that it takes to verify data and the mining.

E-voting

A little definition

The vote (from the Latin votum meaning wish) has been used by humans, for thousands of years, to take communal decisions within a group. Hence, even if characteristics of a vote may change considering the context where used, it has to guarantee some main traits:

Current situation in democratic vote

We will focus on political vote as they generally share the same rules across the world which are: one vote per citizen that equally counts, immutability, privacy, …

Are we there yet?

Legitimacy	In corrupted countries, the legitimacy of the vote is not guaranteed as long as the legal representative can change the result under certain influence Immutability: In some corrupt countries, votes are not really taken into account Verifiability: Even in the most democratic countries, a citizen cannot verify the vote himself (impossible both physically and legally) Security: In corrupt countries, everything can be changed.
Privacy	In most countries, the voter can make his choice in a private space
Speed	Speed is linked to the number of staff hired which is usually around one day to count the vote
Cost	Ex: 2010 UK parliamentary election: £84.6 million for the conduct of the poll (hire staff, print ballots, ...) [20]
Accessibility	Sometimes, queue at the polling station can get really long and demotivate citizen to go voting. Some citizens are too far away from nearest polling station (no vehicle, expatriate)

Do we need a Blockchain?

How would it work?

E-vote would be built on a cryptocurrency Blockchain but instead of transactions of money, it would be transactions of votes represented by tokens.

For a specific election, the organiser of the vote (say the Government) gives one token to each of its voters (the citizens) identified by their public key, exactly like a Bitcoin transaction. Now, each citizen has a wallet with one token to spend for his vote, this would be like his electoral card. Then, the citizen can send this token to a candidate in the same way as previously.

New ideas come from this new kind of poll process. For example, the open source project Democracy.Earth established the concept of Liquid democracy where anyone could delegate its vote by sending his token to a trusted person.

Finally, result would be delivered given the number of tokens earned by each candidate. Thanks to the Blockchain, anyone would be able to inspect the blocks and hence verify the legitimacy of the vote.

Does it do a better job?

Legitimacy	Immutability: Transactions on the Blockchain are immutable Verifiability: Blocks of transactions can be inspected and verified by anyone (public blockchain) Security: What could go wrong in the process? The government creates the tokens with a special id on each: if a hacker tries to create new ones, their id won't be valid at the end They are sent to all citizen's public key registered on the electoral list Wallets on software can be hacked but not if they are stored on hardware like a usb key (can still be stolen physically) A 51% attack (when more than half of the people in a system turn against the system making it worthless due to so many involved in the mining process) is practically impossible: cannot hack so many computers in such a short duration. Candidates receive tokens and the government (end everyone else) can check if a token's id corresponds to an id registered and if it sent by the same citizen's key as the one it was sent to
Privacy	Because everything is public, the privacy of a voter is threatened. Solution: A voter is only represented by a public id but no one knows the identity behind. But a new issue arises: anyone can still see an id's vote. This could lead to corruption of votes. For example, someone could be payed to vote for a special candidate and his task could be verified if he gives his public id.
Speed	Speed could be an issue because every transaction needs to be approved by a miner with a proof of work (Bitcoin 7 operations/s << Visa network: average of 5000 transactions/s). However, this could be fixed by changing the parameters of the proof of work, given the number of voters.
Cost	Cost: only the remuneration of miners (could be 10 cents by vote approved for example: given the UK population in 2010, the parliamentary election would have only cost £6,277 million (more than 10 times less than actual price!))
Accessibility	Voters will only need to have a device connected to the internet which most of the people do nowadays and know how to use it. For the minority who wouldn't have one, they could be some polling station furnished with computers and helpers if needed

A decentralised Democracy?

With the Blockchain, democratic vote could evolve even more by becoming entirely decentralised. This idea is promoted by start-ups such as FollowMyVote or Democracy.Earth. [14]

It would work the same as before but without a centralised identity verification (the government). This time, a citizen registers on the network and need to be approved by miners. This could be done by a digital identity proof published on the Blockchain such as the video of ourselves once a year (See Digital Identity). Hence miners can check previous videos (immutable) and approve today's identity of the user. Once logged, the voter is assigned a token. Finally, everyone can check the matching between an identity, his token and the destination to a candidate.
As such, People would live in a decentralised society that would be surely more secured and true but where the trust in public institution

So do we need a Blockchain ?

Finally, we can deduce that e-voting would be nicely built with a blockchain as it needs to store a record of votes to be verifiable, has multiple writers and we can’t always trust a centralised organiser (a corrupted government for example).
However, it raises a big issue on privacy because there is a potential risk to make a link between a vote record and the voter. In that concern, we could make the ledger more private (like Monero) but this reduce, on the opposite, the verifiability of the poll, hence the legitimacy itself.

Blockchain in Healthcare

In today’s healthcare sector, there are often issues such as loss of data or insecure data which can have negative consequences on the lives of many patients if left unsolved. Some medical systems are not even digitized and there is a lack of common architecture and standards among the different medical systems that exist today whether it be nationally or internationally.

Would a blockchain be useful in the healthcare industry and what would it bring?

Sharing medical data in a secure way

A major problem the healthcare sector faces is the secure sharing of personal data such as diagnosis or blood results. This could be from a professional to a patient or from a professional to another professional. We need to ensure that the data sent is not intercepted and sent to the correct recipient. A permissioned blockchain would allow a safe structure between different parties which may be entitled to pull data from it. The idea would be that the only piece of data held on the blockchain is a pointer to another standard database which contains all the medical records such as X-rays, history, previous diseases. This is what we call off-chain data. The pros of this solution is that we can store data of any format or size. Medical data is quite heavy, and if stored directly on the blockchain, then block processing speeds would decreases and this presents challenges to scaling the system. However, using this solution means that data is not directly visible.[24] The patient itself would have to allow access to someone part of the blockchain by giving his private key or by providing a digital signature which depends on the private key. This means patients have complete ownership of their medical records. The doctor, on the other hand, would then only grant access to test results to his patient and any other professional who needs it. This better sharing of data would allow a higher probability of accurate diagnoses, more effective treatments since even people abroad who need to receive treatment can grant access to foreign practitioners. Blockchain would then allow various stakeholders in the healthcare chain to share access to their networks without compromising data security and integrity, by allowing them to track data provenance. Data is not sent back and forth between hospitals, pharmacies.[25]

The data is safer

Another issue medical organisations face such as the NHS is that any nurse or doctor logged on the system has access to your personal data. This leads to two issues. First, any entry somewhere in the system by a hacker compromises the data of everybody on the system and not just the professional’s patients. Furthermore, personal data can be read by any professional, regardless of them overseeing you. In March 2017 the NHS suffered a severe security breach which caused all its system to be compromised. GPs, and doctors were unable to access patient data putting lives at risk. Since 2011, there has been 700 000 pieces of medical documents lost, misplaced. [26]Your medical records could be worth more than your credit cards nowadays. That is why if a blockchain was implemented only patients could allow access to their records and as a result enhanced security would prevent hackers from stealing your data. Furthermore, a blockchain is difficult to hack because each node has a timestamp. Hence, a hacker would have to take control of all the copies of the node distributed to each participant which is very difficult. This leads to another problem which is what happens in case of an emergency. How can doctors access your data while you are unconscious? The solution to this could be the use of wearables and connected objects which would contain critical data such as allergies, chronic diseases, emergency number and blood type. In the case of an emergency, professional would scan your wearable to retrieve information.[27]

A correct medical record

They are also concerns regarding the fact that multiples entities hold different versions of a medical record. How can we decide which one is correct? In this case the blockchain would be very useful because only valid medical record would be stored on the blockchain and hence all information on the blockchain is correct. This could be made possible thanks to smart contracts. The idea would be to store on the blockchain as well pointers, standardized data such as the date of birth, gender, immunizations and procedures and any other data which is relatively small. Once it is available, data fields can be created in a smart contract to employ rules for storing data on the blockchain as well as making sure there is an approval before storing. On every patient interaction, the healthcare organizations will pass information to the smart contract. The smart contract can validate the data fields and it will then direct the storage to the blockchain if the data is valid. For example the smart contract could require all data fields, or that a specific field contains a particular numerical value.[24] The blockchain would always up to date. As the data on the blockchain should be immutable any attempt to modify it would be visible. The chain here would in effect be the entire history of a patient medical records. The miners, in effect, could be medical researchers or healthcare professionals. Their expertise would be useful in validating different diagnosis. When validated, a new pointer can be added to the blockchain. In exchange, they could be rewarded by aggregated and anonymised data from patients records for further research studies as long as the patients have agreed.[28] This encourages medical progress.

So do we need a blockchain?

A blockchain would allow a standardized safe data structure for the healthcare sector with increased benefits for patients, better diagnosis and progress for medicine. Patients are in control of their data, can share it securely, while their new data can be stored. Indeed, blockchain technology presents numerous opportunities for healthcare. Nevertheless, it is not fully mature and cannot be applied immediately. There are different challenges to solve such as where to find the computing power to run the blockchain, what data exactly is to be stored, will there be an agreement between different medical organisations which work differently. Furthermore can we declare that there is a problem of trust between different medical organisations, I am not so sure.