In a keynote speech given at the Conference in the first half of January 2019, Ethereum co-creator and blockchain celebrity Vitalik Buterin made yet another of his controversial and eyebrow-raising claims.
When an audience member asked him about transactions per second capacity of the upcoming Casper CBC technology, Buterin quickly cut off the audience member before delivering an audience-rousing statement on how many misunderstand and incorrectly prioritize the need for a high TPS rate.
In this statement, he said that projects that employ alternative algorithms to achieve absurdly high TPS rates are all a “centralized pile of trash.”
Cheap, Fast, or Secure – Choose Two
Before we get too deep into the technical analysis of the upcoming CBC Casper developments, we need to discuss what’s really at stake here. Also, why does Buterin feel that projects that overemphasize TPS (transactions per second) are trying to capitalize on buzzwords as opposed to stable and truly decentralized technology?
For those unfamiliar, TPS refers to the number of transactions that a cryptocurrency network can process every second. Bitcoin is infamous for having an unusually low TPS by itself. This is caused by the incredibly high demand on the network that leads to sometimes parabolically increasing transaction fees.
Other, newer networks have higher TPS rates, but could perhaps lack the robust network security that bitcoin boasts. For example, despite bitcoin having a comparatively low TPS, it has, to this day, never suffered from a 51% attack, a time warp attack, or any other primary attack type. Compare this to younger and more nimble projects like Verge which have been successfully attacked multiple times. That’s not to say that a high TPS automatically makes a project less secure. Instead, it’s a reminder that just having a high TPS is not enough to build a reliable, safe, and trustworthy cryptocurrency network.
Addressing the TPS problem is a sore point among crypto enthusiasts. The debates surrounding this can become so unpleasant that virtual civil wars have been launched over it. The classic example being Bitcoin Cash splitting off from bitcoin, primarily under the banner of having higher TPS and lower transaction fees. Before that, a similar break-off occurred with Litecoin (although it was not contentious as Litecoin took on a completely different name and identity, as opposed to claiming to be the true bitcoin).
“Centralized Pile of Garbage”?
The core component of Buterin’s claim at the conference revolves around poorly arranged priorities. According to Buterin, there are many ways in which a high TPS can be achieved. The problem is that not all of these methods are created equal. Buterin specifically called out algorithms like VFT, by claiming that the only way these projects are able to achieve their boasted TPS is by centralization.
Buterin’s team is moving in the opposite direction. Instead of striving for the highest TPS rate possible, their goal is to focus on two things. First, how difficult it is to censor or make blocks revert, and how long does it take to make transactions safe. According to Buterin, these two concerns are what matter, not a sky-high TPS rate.
Specifically, Buterin said:
“When a blockchain project claims we can do 3500 TPS because we have a different algorithm, what we really mean is we are a centralized pile of trash that only works because we have seven nodes running the entire thing.”
In other words, Buterin states that these projects can only achieve their TPS because they rely on what is essentially a central point of failure. This network design methodology is the antithesis to how cryptocurrency networks are supposed to work.
Think of it this way. Each bitcoin block is only 1 MB. By today’s standards, that amount of data is microscopic. If we were to host all of the bitcoin network’s processing power onto, say, the Amazon AWS cloud service, it could easily process billions or even trillions of transactions per second.
So why don’t we do this? The answer is simple. Bitcoin was designed to be decentralized, which means that it is immune to censorship and attempts by outside parties (think governments, police, etc.) to shut it down. If all of the bitcoin network transactions were operated on a cloud service owned by a company, it could be taken down in seconds by the request of a government entity to which the corporation that runs the network would be forced to comply with.
Essentially what we have here is a trade-off. The bitcoin network sacrifices the incredibly high speeds and network throughput capability of a highly developed centralized system in exchange for a much slower one that is immune to censorship and takedowns.
The main element of Buterin’s claim is that these unnamed projects he is referring to are essentially going this route. They are choosing centralization and speed over decentralization and security.
To support this, he said: “The purpose of consensus algorithms is not to make a blockchain fast … [it’s] to keep a blockchain safe.”
All About Casper CBC
How can a project achieve a high TPS while still remaining decentralized and secure? Buterin stated that instead of relying on tech dedicated to maxing out TPS, higher TPS rates can be achieved by using technologies such as sharding and other layer-2 technologies (other examples being Lightning Network or Plasma).
One such technology that the Ethereum Foundation is working on is known as Casper CBC. The CBC stands for correct-by-construction. It relates to a set of new protocols that will impact how consensus is achieved on the upcoming Casper implementation of Ethereum.
Ethereum R&D member Aditya Asgaonkar describes CBC like this: “It is called correct-by-construction because correctness of new protocols is guaranteed by their construction, which is by incrementally defining an abstract (and provably correct) protocol further.”
The Casper implementation has been anticipated for quite some time now, and promises to move the network away from arguably wasteful proof-of-work consensus – meaning mining on ASIC devices that consume large amounts of power and resources, and converting it into the much less energy intensive proof-of-stake model.
Proof-of-stake models require many more advanced techniques to achieve a high TPS rate, but it is entirely possible as long as the developers find a way to do so without sacrificing the security of the network.
Buterin the Spitfire
Calling out other projects, technologies, and people is something that Buterin has become somewhat famous for. In the past when he has made claims like this, the response is usually divided between those that agree with him wholeheartedly and those who find his rhetoric shortsighted or incomplete. In this case, Buterin did not mention any specific projects, and so we are only left to make educated guesses on what networks he is referring to.
In the short term, however, Buterin’s comments are still mostly bluster without backing. Some information has been made public regarding the development of Casper and CBC, but the general public has still not been given full access to Casper. Therefore we don’t know if it can accomplish everything that Buterin has promised.
Perhaps we will know more details on Casper, and be able to determine directly how much truth is contained within Buterin’s emotionally charged declaration before the end of the year.