Ethereum’s Training Wheels

The temporary role of centralization in blockchain adoption

“Learning” how to use blockchain technology won’t be a simple task for everyone, but the same can be said about riding a bike. Standard bicycles are two-wheeled vehicles. When you add two more, you lose it’s versatility, speed, and purpose. But without training wheels available to teach people the skill of riding, we’d have far less bike riders in the world.

Without the points of centralization in blockchain (i.e. exchanges, mining, and blockchain access API’s) we would have seen less early adoption of the tech and less “crypto folk” utilizing and championing it. But if we continued on the path of centralization and assumed altruism, we would never become fully decentralized. It’s important to identify and recognize these centralized crutches that, just like training wheels, once we’ve outgrown them, will need to be abandoned. So let’s start by talking about the area of blockchain centralization that brings out the most anger in some. The exchange.

Centralized Exchanges

“I definitely hope centralized exchanges go burn in hell as much as possible.”

-Vitalik Buterin, Founder of Ethereum

An eternal bath in hellfire may be a tad bit extreme, but the ideal version of blockchain will definitely not include centralized exchanges. But if that’s the case, and the brilliant founder of Ethereum is so vehemently opposed to them, why have centralized exchanges always dominated the market?

For starters, the technology for decentralized exchanges (DEX) just isn’t quite there yet. DEX’s are still in their infancy. They can’t yet handle the same volume and velocity of transactional throughput that centralized exchanges can. Centralized exchanges certainly helped get us to the current level of adoption in the industry. Without them, the current volume of trades would have never been reached, and adoption would have increased at a much slower rate. This, combined with centralized exchanges being the only readily-accessible gateway to crypto from fiat currencies, the higher functionality of exchanges, and a centralized exchange’s ability to insure users’ money all contributed to why centralized exchanges dominated the beginning of blockchain.

But DEX’s are slowly gaining momentum. More and more money is being poured into them, and there are already several competing DEXs. Although DEXs have their own share of security woes, switching away from centralized exchanges should result in less hacks, less infrastructure downtime, more privacy for users, and greater resistance to censorship or other access restrictions. Centralized exchanges also just have too much power. For instance, they can influence the popularity of different currencies.

I’ve attended a few conferences recently and what’s clear is that many centralized exchanges are focused on capital gains. They are opportunists who are profiting off the current demand for crypto. The employees of centralized exchanges I met are not interested in advancing the technology. Decentralized exchanges are on the opposite end of the spectrum. They are focused on the true value of decentralization, yet unable to keep up with the sheer size of market demands. DEX’s aren’t thriving businesses yet, but they’re fortunate enough to have the financial support of investors and the community alike.

Centralized exchanges profit every time there’s a transaction. It’s in their best interest to foster volatility in the market. Centralized exchanges are far too focused on profits. In fact, the main reason Vitalik wants them to burn in hell is that there is “no reason some projects need to pay $10 to $15 million in listing fees to let people trade their tokens on centralized exchanges.” Centralized exchanges may be the training wheels we need right now for adoption, but we should never give up on DEX’s or at least a semi-centralized hybrid model.

Exchanges may have most of the limelight, but there is an equally dangerous yet less understood area of centralization we need to keep in mind. Let’s talk about how everyone is getting access to the blockchain.

Centralized Blockchain Connectivity

At the time of writing, there are currently 14,252 nodes connected to the Ethereum Mainnet. Compare this to the 40,487,243 unique addresses on the Ethereum Mainnet. Now it’s important to note that not all addresses equate to one user. A user can have multiple addresses, and even smart contracts themselves carry an address. But even if only 1% of addresses represent actual distinct users, you would find that over 95% of blockchain users are relying on someone else’s node for blockchain access.

This is a serious problem. You use a blockchain to get and interact with data in a decentralized and trustless manner. However, a vast majority of users are trusting someone else’s node to altruistically interact with the blockchain on their behalf. Currently, the only trustless way to interact with the blockchain is by syncing up your own node and connecting to the blockchain yourself. But people just aren’t doing that.

Awareness of this issue has been pretty low up until recently. There are a few different Remote API’s offering blockchain access as a service, most notably Blockcypher and Infura. Infura dominates the market and is now processing 10 billion requests a day. 10 billion requests for decentralized data are being passed through a centralized pipeline. When you keep the training wheels on the bike too long, at the very least, you aren’t getting the most out of the vehicle. Worst case, you never learn to ride without them. When you interact with decentralized data through a centralized pipeline, you’re no longer decentralized. If we rely on centralized pipelines, we’ll never be decentralized. Does this make Infura evil? Absolutely not. We could never have gotten where we are without them.

The CTO of Colendi, in a recent blog post, asked: ”Infura is a great tool but doesn’t it clash with the term of decentralization?”

Infura has made it incredibly easy for people to access ethereum-based services. It’s not a perfect version of blockchain services, but like with centralized exchanges, Infura provided necessary training wheels that helped the surge in adoption we’ve been seeing recently.

As the network continues to grow, this dependence on Infura is worrying some. A journalist for the Crypto Daily Gazette pointed out that ”This dominance of Infura presents a situation of centralization, as this company is the primary provider of nodes in the Ethereum (ETH) network, which has thousands of developers and decentralized applications.”

But let’s walk through a real life example of how using something like Infura can seriously undermine the value of Ethereum. Let’s take the Ethereum Naming Service (ENS) for example. ENS is great for decentralization. It allows users to tie names to addresses or contracts. This is just like DNS with standard web browsing. When you want to find out information online you go to “google.com” not an IP address like http://74.125.224.72. In the same way, ENS allows you to send ether to “alice.eth” instead of to 0x06012c8cf97bead5deae237070f9587f8e7a266d.

This matters for a couple reasons. First, it makes everything easier. It’s simply better UX to allow people to send money to a simple word-based address than a string of hexadecimal characters. Especially when an address is 40 characters (42 if you count the obligatory `0x`). Secondly, it reduces human error. It’s easy to see with a glance of an eye that “goofle.com” is not the same as “google.com”. But try and spot the differences between these two:
(there’s 3)

A. 0x06012c8cf97bead5deae287070f9587f8e7a266d

B. 0x060t2c8cf97bead5deae237070g9587f8e7a266d

It’s not impossible. You may even have found them at first glance, but that won’t be the case every time. It’s good to reduce the capability for human error. This matters more than you may think. One of the most common mechanisms of a phishing attack is to distribute incorrect addresses via fake or hacked accounts or websites. A lot of money has been intercepted this way.

So why does Infura undermine the value of ENS? For the same reasons blockchain is good.

Julian Martinez, of Quantstamp, accurately points out that ”We are trusting that Infura is giving us correct information about the blockchain. This also seems like a central point of failure.” This is absolutely a central point of failure. And the size of their operation has made things even more dangerous.

There are three main situations that make using Infura to access blockchain systems like ENS dangerous. First, Infura could act maliciously. They have the ability to give out false data or send their own address when you ask for Alice’s. Given the risk this would pose to Infura’s position in the market, this is an unlikely scenario. If we trust Infura to be a rational actor, we can negate this possibility. That being said, it’s not about being trustworthy, it’s about being trustless.

Hanlon’s Razor states that we should “Never attribute to malice that which is adequately explained by stupidity.” That’s the second problem here. Someone at Infura may not be evil, but dealing with complex data infrastructure is no easy task. Any human error on their end results in a problem for the network at large. They could expose a vulnerability or do something that results in infrastructure downtime. Infrastructure downtime from Infura would impede an overwhelming share of decentralized services on Ethereum.

But the greatest concern is the giant target painted on Infura’s back. As a central point of failure for the network, and as an entity with the ability to “do too much”, they are the ideal target for hacking (including social engineering). If Infura were compromised, a significant amount of damage could be done. Ideally, there should exist simple and convenient means of blockchain access which do not introduce glaring single points of failure.

Blockchain access is one of the most devastating areas of centralization when you consider how widespread it is. There are several DEX’s being worked on, but there doesn’t seem to be much of a community attempt to tackle the problem of decentralized blockchain access.

An arguably smaller problem, but one that surely gets more attention, is centralization in mining.

Centralized Mining

Centralized mining isn’t the same threat as a centralized blockchain, but it’s still less than ideal.

It’s important to note than even if mining is centralized, miners can’t simply create a transaction out of thin air. They don’t have complete control and can’t violate the fundamental rules of the system. For example, If Alice runs the biggest mining pool and has a balance of 3 ETH, she still cannot execute a transaction sending 5 ETH.

Also, basic Proof-of-Stake game theory applies to any group that has a majority mining share. A majority of hash power on the bitcoin blockchain comes from miners made by Bitmain. This shows a clear centralization of manufacturing. But if Bitmain is heavily invested in the bitcoin blockchain through their business, it will likely never be in their financial interest to do anything malicious. Their stake is too high.

Hash power centralization is another story. If some entity controls a majority of the hashing power in a blockchain, they can control what to include and what not to include in the ledger. The “target state” is up to the controlling power. The controlling power could reorder recent history and censor parts of what gets recorded in the blockchain. This undermines the blockchain’s ability to function as a reliable and open marketplace. Fortunately, decentralized mining pools are possible and could remove many of the downsides of centralized pools.

Centralized mining also amplifies perverse incentives around protocol updates and improvements. Proof-of-stake is clearly not in the best interest of larger centralized mining pools. Switching to proof-of-stake is seen as a threat to profits for many of these pools. This creates an incentive to resist protocol changes which reduce or eliminate reliance on classical mining strategies. Ethereum developers have been attempting to address this issue by adopting a gradual transition to proof-of-stake, as well as baking so-called “difficulty bombs” into the protocol which serve to force periodic protocol updates.

Centralized mining may not pose the same class of risks that other elements of centralization do, but it is still undermines the promise of strong decentralization. Jimmy Song of Bitcoin Tech Talk, goes into much more detail about mining centralization here.

There are two other areas of centralization that we should mention: centralized figureheads and development. These issues are harder to codify, and are typically regarded as less problematic than centralization of infrastructure, but they do merit discussion.

Centralized Figure Heads / Vitalik

Across blockchains, there are influencers. Bitcoin Cash has Roger Ver, Ethereum has Vitalik Buterin.

It’s important to note that, without Vitalik, Ethereum would be nowhere. Ethereum is a global project, but it’s Vitalk’s baby. Through his leadership, Ethereum has navigated through the muddy waters of creating a decentralized ecosystem. Vitalik shares a good bit of responsibility for the hard fork that helped recover funds from the DAO hack. He’s a big reason why Ethereum plans to switch to proof-of-stake and is at the forefront of scaling the network.

Let’s compare this to Bitcoin for a moment. Satoshi Nakamoto has remained anonymous and withdrawn from the community. There is no central leader for Bitcoin. When Vitalik proposed the foundations of Ethereum to the Bitcoin community, they rejected him. In fact, Bitcoin has been significantly less ambitious to changes than Ethereum. As a result, Bitcoin may seem more consistent, but it is less flexible. The lack of ideological leadership in the Bitcoin community is likely a contributor to its comparative stagnation, and stagnation is dangerous in tech.

Last year, Vitalik told Vice, “…there are people who oppose me on technical matters and half the time I’m right, and sometimes they’re right, but if someone opposes me socially, often there are many people who come to my defence and that seems like a fairly stable coalition. But, there shouldn’t be a coalition around any one person. It should be a coalition around the protocol itself.”

When asked about passing on the mantle, he said, “That might happen in 5 years.”

There is a cult-like status around Vitalik. Check out fitvitalik. I’m still not sure if this is a joke or a scam, but it appears there are people genuinely raising money to help Vitalik get more healthy. They’ve raised over $10,000 USD. There are also excellent Vitalik memes and works of art.

Vitalik once said that, “If someone puts a gun to my head and tells me to write a hard-fork patch, I will definitely write it. If I publish a patch to delete a bunch of accounts, how many people here would download and install the update and switch to that chain? I see relatively few raised hands. This is called decentralization.”

Maybe not the majority, but some people would install a malicious update if Vitalik told them too. It’s important to clarify that his power isn’t executive. Vitalik holds cultural power more than anything else; it’s closer to a technically-proficient celebrity than a commander-in-chief. Ideological influencers are, likely, an inevitable product of the human condition. Effective checks and balances are probably more appropriate than decentralization in the strictest sense.

Centralization in Development

There are a few complaints looming around that Ethereum development is far too centralized. A study by University College London researchers revealed some statistics on the matter. They studied bundles of changes a developer submits to the codebase of a blockchain. They found that, “7 percent of all the files in the Bitcoin Core software were written by one developer, while about 20 percent to ethereum were written by a single coder.” A fifth of changes submitted to the Ethereum codebase were done by a single developer. Does that give this developer too much power? Not necessarily.

There’s a big difference between few developers writing the code and few developers reviewing (and understanding) the code. If a codebase is open, and reviewed by developers with diverse incentives, who originally wrote any given line is an implementation detail.

It looks better on paper if updates come from a variety of developers, but as long as blockchain software continues to be scrutinized, threats of developer centralization are negligible.

Conclusion

Many points of centralization persist in Ethereum and across the broader blockchain ecosystem. We need to be aware of the training wheels on both Ethereum and blockchain at large. Whether it’s exchanges, blockchain access API’s, mining, or the need for awareness of over-centralized areas of power.

When you’re trying to ride the decentralized path, continuing to use centralized dependencies ruins the purpose. Training wheels are meant to be a temporary solution to help people learn. As blockchain is in its infancy, it’s okay to lean on these crutches and benefit from these centralized aids. When the time comes, once the network matures, we will have to move on.

A common strategy parents use to wean a child off of training wheels is raising the training wheels up by an inch on each side. In the same manner, it is likely that we will need to wean away from centralization by first adopting semi-centralized models. We should accept this as a reasonable strategy on our iterative approach towards the goal of true decentralization.

No one can say with certainty when blockchain will reach adulthood. There is far too much uncertainty in the space. What we can say with certainty is that the technology is growing and evolving. A common piece of advice shared with parents is to wait until the child is ready before removing their training wheels. We’ll get there one day, but today, Ethereum just isn’t ready yet.

Contact/Connect with us at:

Twitter || Facebook || Telegram || Website

DISCLAIMER: The content provided on this site is opinion and commentary on topics related to the blockchain universe. IT IS NOT INTENDED TO BE NOR SHOULD IT BE RELIED ON BY YOU FOR ANY REASON AND IS PROVIDED “AS IS” WITH NO WARRANTIES OF ANY KIND. You are responsible for your own decisions and for properly analyzing and verifying any content.