Dean
Latest posts by Dean (see all)

yoda_path_darksideDecentralization is the single most important characteristic of Bitcoin. It is the core property of the network on which its entire value is based, regardless of whether you measure that value in economic or social terms. This is what Bitcoin is: a decentralized currency. Everything else, every other property or feature, every other advantage that it can offer, is fed by this singular beating heart.

But decentralization is not a simple black and white thing. It is not a simple matter of a tick or cross in a box as to whether Bitcoin is decentralized or not. There are varying degrees of centralization and decentralization. Crucially, it is important to recognize that Bitcoin is not entirely decentralized, and that the degree to which it is can vary over time and be influenced by a wide range of different factors.

In this article I will look at the different forms of centralization which affect Bitcoin and why they are a problem. In doing so, I don’t want the reader to be left with the false impression that I believe any of these things to be a fatal problem. More decentralized is better, more centralized is worse, but we will always be somewhere along the continuum between the two – 100% decentralization is probably not possible and arguably wouldn’t be desirable anyway. This is because there are always trade-offs, and the closer you get towards complete and unblemished decentralization the smaller the benefits will be with each step, and the larger the costs will be. Having said that, I do believe that the degree of various forms of centralization afflicting the network at the moment is a problem that needs to be addressed.

The Centralization of Bitcoin

Centralizing forces acting upon can be grouped into two primary categories and multiple sub-categories. This classification system is entirely of my own devising and created purely for the purposes of this article, so please don’t mistake this for some kind of objective system or conventionally accepted wisdom.

Fundamental Threats to Decentralization

This category includes everything which objectively impacts the degree of centralization or otherwise of the network itself and its fundamental components. I have identified three types of centralization within this category.

Mining Centralization

This is usually considered to be the most dangerous form of centralization. It is probably the best known as well, although I would argue not always the most understood.

The domination of mining by an increasingly small number of entities is a problem for a multitude of reasons, the most famous of which is the threat of the so-called ‘51% attack’. In plain English, this means that any entity which is able to control greater than half of the computing power put into mining would be able to ‘double spend’ coins – using coins to sell for fiat or make a purchase, and then reclaiming possession of those same coins.

This may not actually be as a big a risk as it is sometimes made out to be, since an attack like this would negatively affect the value of Bitcoin, which in turn would reduce the value of the mining business carrying out the attack. This means that there would be a big cost for the attacker to pay, and at least part of that cost may be paid even if the attack was unsuccessful. It is difficult to imagine many situations in which the potential rewards for this would be enough to offset this cost.

On the other hand, it may also not be as black and white as some people think. Everybody understands that if a single entity controls over half the mining power this is a problem, and as a result various pressures generally prevent this from happening (or at least correct the situation fairly rapidly if it does). But I would like to suggest that even if it takes two or three entities combined to reach 51%, this already subjects the network significant part of the risk associated with this attack. The reason or that is simply because, although a mining business itself may not willingly attack the network, their power may be co-opted by another entity which does not stand to lose financially from a crash in the value of BTC. In fact, if they were to take out a short sell on the price of BTC, they may be able to make more money from the drop in price than from the double spend itself. This attacker may gain access to mining power through hacking, violence / threats of violence, bribery of employees etc. rather than through actually owning a major mining business themselves. Although it is somewhat more difficult to do this to two or three companies simultaneously than to do it to a single company, it is only two or three times more difficult and perhaps even less (e.g. the same software may be used as the basis for breaching the security of more than one organisation). So it is not so simple as just thinking that as long as no single person controls >50% then this attack is not an issue – it is more accurate to think that with a strongly decentralized network the risks are vanishingly small, but the more centralized mining becomes the more this issue becomes a threat.

And the problems caused by the centralization of mining go beyond just this one potential but unlikely attack.

It is the miners who decide which transaction get to be included in any particular block. Allowing power over this function to be centralized into the hands of a small number of entities may not be a critical security issue, but that does not mean that it isn’t a problem.

The power to decide whose transactions get included into a block means the power to offer an express service to some people or to purposefully delay the transactions of others. Of course a certain amount of prioritization based on fees paid is inevitable – in fact is will be essential for the network to remain secure as the block reward goes down. But I would argue that there is a big difference between this happening via a truly decentralized fee market and it being centrally controlled by a very small number of major players.

Gaining control over this power to select transactions for inclusion in a block allows for deals to be struck giving large corporations priority transaction processing privileges over ordinary users, effectively allowing them to pay wholesale through a partnership deal with one or two of the big players in mining. This is already happening. Some may not see this as a problem, but it does further the advantages of large centralized services running on top of the blockchain over direct use of the blockchain itself, something I will look it more further on in the article. It also distorts the fee market, making it difficult for other users to know how high a transaction fee they need to pay, because they won’t know which transactions are being included due the fees they pay within the transaction and which are being included due to one of these partnership deals. What’s more, the same power can be used to delay transactions – opening up the possibility of attacking competitors by making the network impractical for them to use, or to hinder the use of the network by individuals who have fallen out of favour for some reason. As smaller miners are not in a position to negotiate such deals, this also serves to enhance the advantages of the large players leading to even greater centralization.

Whoever controls mining also controls the protocol. Its is primarily the miners who get to decide which, if any, changes can be made to the code which runs Bitcoin. If the majority of the hashing power decides for or against a change it is almost impossible for other users of the network to go against this decision – because by doing so they would create a fork which could easily be attacked by these miners. Centralized control of mining is therefore de facto centralized control of Bitcoin itself.

There are three distinct but related forms of mining centralization.

Pools

The use of pools is essential for most miners to reduce ‘variance’, which is the varying amount of time a miner has to wait before earning a reward. A solo miner may have to wait a long time before earning anything – potentially even so long that their equipment has already become unprofitable before they even earn anything at all. Belonging to a large mining pool not only means that the miner doesn’t have to worry about running their own full node, it means that they can start earning rewards for their contribution pretty much straight away, and continue earning those rewards according to a reliable, predictable schedule.

But this means that a small number of pools are able to take control over a large proportion of the network’s entire hashing power. All of the power that miners have over the network is effectively given over to these large pools.

In the past there has been at least one case in which a single pool gained over half of the network’s hash power. At the time of writing this article, the top two pools control precisely 50%. As I explained above, I am of the opinion that this is a serious problem.

Economic Centralization

This refers to the domination of mining by a small number of economic entities, generally as a result of economies of scale. These economic entities may be the owners / operators of large mining farms, or they could be the producers of ASIC chips who undoubtedly hold a great deal of sway (if not actual ownership) over those same mining farms who they are able to supply with chips that aren’t available to the general public. It is more difficult to gauge the size of the big players in this regard than in is to monitor the size of the large pools.

Geographic Centralization

It is common knowledge that mining today is dominated by Chinese companies. This creates an additional set of problems, not because there is anything particularly sinister about Chinese people, but because centralization within any single legal jurisdiction is a problem.

Bitcoin is subject to control by the miners. Businesses are subject to control by the government of the country they are in, through legal pressure (which, I might add, may be applied in secret without the general public knowing about it). Therefore if all of the mining power is located in a specific country then the government of that country may effectively be able to control the Bitcoin network. The secretive and authoritarian nature of China’s communist government only heightens this risk.

Not only that, but actions taken by the Chinese government may have a disastrous effect on the network, even without them launching an actual attack. What if China changed its laws to make mining illegal? Not only would the hash rate drop massively and possibly overnight, which could create a wide range of problems (until difficulty caught up with the change) but there would be a lot of ASICs sitting around with little legal value to their owners, creating a significant risk that they may end up being used for a 51% attack.

Because governments do not have the same economic incentives not to attack or otherwise harm the network and its value, which generally help to keep mining companies honest, I personally consider this to be the most dangerous form of centralization.

Other risks include a loss in the internet connection between China and the rest of the world, due to natural disaster, war, terrorism, or the government cutting off the internet in the event of widespread social unrest.

Diversity of Node Ownership

The next concern regarding Bitcoin’s level of decentralization relates to full nodes. These nodes provide vital functions, including relaying transactions and acting as auditors to validate the blockchain.

There are some concerns that the number of full nodes is falling, and that increasing bandwidth and storage costs as the network grows will serve to accelerate this trend, creating another troubling form of centralization.

The number of full nodes is often taken as a measure of this, but that is not quite correct. In actual fact, it is the diversity of node ownership which is most important. If there were 10,000 nodes around the world, but they were all owned / controlled by two multinational corporations, then that wouldn’t be very decentralized. The key is how difficult and costly it is to set up and run a node, and how may independent people or companies have enough incentive to do it.

As I will explain in a future post, I do not personally consider this to be anywhere near as important as mining centralization, but it is certainly not something that can be ignored. It is also something that many people can help out with by running a node themselves at home! Take a look at Bitseed and BitcoinMini for the easiest way to do this.

Centralization of Development

Bitcoin is open source software. Anybody can contribute suggested improvements to the code. Multiple competing versions can and do exist alongside each other as well, and to a certain extent users can choose which version they want to run – although its not quite as simple as that due to the economic issues and complexities around creating blockchain forks, as well as the fact that it is really the miners who get to choose rather than users, because any fork not supported by the majority of miners would have a severe security issue.

But there is still a small group of programmers who effectively control development, who possess the ‘alert key’ to advise users when to upgrade and what to upgrade to.

This is causing controversy at the moment as because the core developers are currently being paid by a company which is developing off-chain scaling solutions. Since one might think that the core developers’ main job at the moment is to develop on-chain scaling solutions, this is at the very least a conflict of interests, and has led to the proliferation of speculation that there may in fact be a conspiracy to hinder the network’s ability to scale in order to serve the economic interests of this company. Regardless of whether you agree that this is happening now, the mere possibility that it could happen is troubling in regards to Murphy’s Law: ‘if something bad can happen, it will happen’; maybe not today, maybe not tomorrow, but at some point it will definitely happen.

Additional Threats to Decentralization

This section includes those centralizing forces and tendencies which do not impact the fundamental nature and functioning of the network, but nevertheless do have a practical impact.

Corporate Services vs Direct Blockchain Access

This refers to the degree to which centralized services built on top of the blockchain are able to offer advantages in comparison to direct use of the decentralized network by end users.

The issue of on-chain vs off-chain scaling touches on this issue, as does the issue of powerful miners cutting private deals to prioritize transactions from certain companies. Geographic centralization also increases the possibility of government regulation impacting the freedom of users to directly access the blockchain.

Although I do not see this as a major area of concern at the moment I thought it was something worth mentioning as there may be other circumstances in the future in which development choices may determine the competitiveness of direct blockchain access versus mediated access via corporate services. Its also a truism that if you don’t use it you generally lose it, so I would encourage readers to consider carefully which services they choose to make use of.

Media Control

As a somewhat democratic entity Bitcoin has some things in common with politics. And just like politics, the ownership / control of the media is highly significant. The opinion of the average person, who does not have the time to thoroughly research and delve into the details themselves, is heavily influence by the media sources they rely on. This has come to the fore in the recent block size debate, with Bitcointalk.org and Reddit/r/bitcoin – the two most popular media locations for digesting and discussing information about bitcoin, censoring and hiding opinions that their moderators don’t agree with by artificially labeling their opponents opinions as referring to an altcoin rather than Bitcoin itself.

Conclusions

Bitcoin’s status as a decentralized currency is not guaranteed. It is something which must be vigilantly maintained.

There are degrees of decentralization, and we are moving in the wrong direction.

I hope that by laying out the main issues as I see them I can in some small way contribute towards a conversation which I believe needs to have greater prominence among the users of Bitcoin.