What is Decentralization anyway?
Decentralization is important, or so we are told. Clearly, there must be more to it than meets the eye for it to deserve the attention that it receives. It’s a matter that merits some further scrutiny and nuance: network decentralization points to the fact that a network does not have a single point of failure, or in other words, that its vital parts are spread out. A traditional company relies much on its centralized decision making to function whereas most modern companies have their decision making more spread out or decentralized; the matrix organization is the extreme example of that. But how exactly can cryptocurrency networks be decentralized or centralized? This rabbit hole goes pretty deep. First let’s talk about the elephant in the room: decentralization-centralization is not a binary matter but a multifaceted scale. A network can be fairly decentralized in one aspect but completely centralized in others. Ethereum’s Vitalik Buterin writes very clearly about software decentralization in his Medium article. He states that:
When people talk about software decentralization, there are actually three separate axes of centralization/decentralization that they may be talking about. While in some cases it is difficult to see how you can have are e one without the other, in general they are quite independent of each other. The axes are as follows:
- Architectural (de)centralization?—?how many physical computers is a system made up of? How many of those computers can it tolerate breaking down at any single time?
- Political (de)centralization?—?how many individuals or organizations ultimately control the computers that the system is made up of?
- Logical (de)centralization— does the interface and data structures that the system presents and maintains look more like a single monolithic object, or an amorphous swarm? One simple heuristic is: if you cut the system in half, including both providers and users, will both halves continue to fully operate as independent units?
He goes on to point the reasons for decentralization as being:
- Fault tolerance— decentralized systems are less likely to fail accidentally because they rely on many separate components that are not likely.
- Attack resistance— decentralized systems are more expensive to attack and destroy or manipulate because they lack sensitive central points that can be attacked at much lower cost than the economic size of the surrounding system.
- Collusion resistance—?it is much harder for participants in decentralized systems to collude to act in ways that benefit them at the expense of other participants, whereas the leaderships of corporations and governments collude in ways that benefit themselves but harm less well-coordinated citizens, customers, employees and the general public all the time.
Buterin ends with a discussion of these three error types and what their importance is for his project, Ethereum, as well as similar projects. An interesting read, for sure.
What this means applied to CryptoNote coins
Some interesting things come out of this. Let’s make a rough analysis here. We can see that CryptoNote blockchains whose function is to serve as sound money are strongly architecturally decentralized. Game theoretically speaking, coins that have mining pools will (ceteris paribus) be less decentralized than those that do not accommodate pools since there is no stringent reason for the network participants to have their own full node in the network. The seed nodes, which are hardcoded node connections in the wallet software of a network, could be seen as a weakness here but their unlikely failure can always be worked around, as long as network participants have a way to communicate with eachother through other channels.
Politically, CrytoNote blockchains are fairly decentralized. While no one controls these networks and the community can ‘fork’ away control from a compromised developer -as happened with Bitmonero- most smaller coins have relatively small teams making design decisions. This is not necessarily a bad thing: firstly, any project worth its weight in salt will be completely open source so anyone can contribute, which is what often happens as a coin gains more exposure. Secondly, not everything has to be crowdsourced because the users don’t always think everything through the way an experienced protocol programmer would. Given the chance, Bitcoin would have increased its block size in 2017 to gain a short term advantage; as things have evolved up until now, the decisions made back then by the developers turned out to be the right ones for Bitcoin’s long term survival and core principles. So while the whole idea of small teams isn’t bad in itself, it does introduce the Bus Factor: what if they get hit by a bus or die in a plane crash?
Logically, CryptoNote blockchains are centralized because the idea is that all the nodes in the network agree on a given state of the system. Especially for a blockchain that wants to be used as sound money, logical centralization if important.
Attacks – a good enough reason to want decentralization
Let’s look at the three proposed attack vectors in the context of CryptoNote coins: fault tolerance, attack resistance and collusion resistance. In its essence this is a matter of risk appetite: smaller projects will have less worth at stake and can try out more radical innovations, which carry more risk. Eventually you want ways that harden a project in a built-in way.
Thanks to the strong architectural decentralization, sound money CryptoNote networks will be very tolerant to hardware faults of the individual participant. Secondly, even from a development standpoint, those with small development teams can offset their potential systemic faults (e.g. a fauly client is pushed out and it is the only implementation) with the open source nature of the projects and adequate communication. It cannot be stressed enough that keeping code closed source in a blockchain project will de facto bring back the trust requirement, which was one of the main reasons to accept all the performance tradeoffs that decentralization brings to the table.
Having a network where no node is ‘more equal’ than other nodes, i.e. a level playing field for all participants, is a good start for a network’s attack resistance. Masternodes and other shenanigans do not have any credibility when sound money is the goal. This is where most CryptoNote coins have some homework, though. Mining pools hold great power over the coins they are mining. The rational miner argument, i.e. that miners would move away as soon as doubts pop up about a pool operator’s trustworthiness, do not seem very convincing if you see that (at the time of writing) some miners keep their machines in pools operated by proven malicious actors. Another argument against the rational mining myth is that, despite the expectation of the miners to spread out their hashing power and prevent any one pool to reach 51% of the network hashrate, several of the larger CryptoNote coins often have mining pools that go far beyond this percentage. These are clear examples of human nature and it is doubtful that there exists a perfect answer to these challenges. However, dark as it may seem now, the future is not overly bleak on this front: an initial attempt at centralizing the reward system while keeping the actual mining decentralized is currently being developed by Nerva community member Nithronium. The concept goes by the name of ‘pond mining’, a term coined by community member Yeuph. A first draft of the technical side of pond mining can be seen in the following write-up. Nithronium – Pond mining: A Peer-to-Peer Trustless Pool Mining System.
Let us then, lastly, assess the collusion resistance of sound money CryptoNote coins. Collusion resistance points to the idea of multiple network participants working together to gain an advantage over others. The weakest point in any of the current CryptoNote implementations is the distribution: new coins are created by mining through the well known Proof-of-Work algorithm. In its most basic implementation, this means that one CPU gets one vote. Of course you want the process to be as fair as possible. Other cryptocurrencies, like Bitcoin also started out as solo CPU mineable only, yet this did change quite fast. After a few years of Bitcoin mining on first CPUs and after that GPUs (Graphics Cards), ASICs (Application Specific Integrated Circuits) appeared on the scene. In stead of having a more or less level playing field where anyone (with a computer) could viably participate, people now had to buy expensive specialized hardware solely to mine Bitcoin. While ASICs may be more energy efficient, their prohibitive cost as well as their high electricity usage and the risk of them being outdesigned by a new generation of ASICs are one of the most centralizing aspects of Bitcoin mining. As we noted earlier in this article, the game theoretical fallacy of the rational miner does not work out in real life situations. The end result here is that ASIC mining throws up strong barriers to entry in a market where you want as many people as possible to participate. Monero, the biggest and most well known of the CryptoNote coins has taken it on itself to fight off potential ASICs that they find on their network. They will only allow for GPUs to mine on their network. CPUs are also welcome, of course, although that is economically unviable. The fight against ASICs is an ongoing fight they will continue until there is no single (cartel of) ASIC manufacturer(s) that milks the market, at which point ASICs should become a less financially restrictive endeavour for the hobbyist user. When such a time comes, if ever, the technology that makes up these devices should also have matured more fully.
Back to the roots.
Mid 2018 a new CryptoNote implementation was released by Aussie developer Angrywasp. It is called Nerva and is somewhat experimental in nature. Nerva has an accelerated emission to see what will happen to a CryptoNote coin’s economics once the tail emission phase has been reached. The other goal it has set itself is to be only mineable through CPUs and without centralized mining pools as we know them. While the first few versions of the adjusted CryptoNight algorithm used in its PoW consensus were not exactly foolproof resistant to GPUs or pools, the newer revisions have proven themselves quite robust and so far no GPUs or mining pools have been detected on the network. These traits, CPU only and the lack of pools, make it so that the Nerva network has an exceptional amount of full nodes, given its total userbase, while the CPU mining side of things makes it much more energy efficient than even the latest generation of ASIC devices.
We can look at it from many technical angles and to some extent we did in this article. On a philosophical basis, though, decentralization in networks like the CryptoNote currencies is simply about making sure no single party has more to say than others. Nobody wants his or her money to be controlled by others: not by banksters, not by mining cartels, not by ASIC manufacturers. It may be a term that is loosely thrown around but decentralization actually does matter greatly.