Google Cloud offers Cypherium blockchain. But what are you actually getting?
When you want to compare blockchains at the ground level, it helps to go back to basics.
Cypherium is now available on Google Cloud, presenting a new enterprise-grade blockchain option through Google Cloud.
So, why would a company choose Cypherium? What exactly does a company get when it picks a blockchain protocol off a shelf, and what would they be looking at when comparing the different options available?
Back to basics
Going back to basics, blockchain is basically a decision-making system. As per the original Byzantine general analogy, it's a way for the separate generals of the Byzantine army to agree on a coordinated plan of action when attacking an enemy city.
They don't know what the plan is going to be, but they do know that they're all going to be agreeing on a single course of action, so each general is sending messengers to the other's camps detailing their preferred plan of attack. Unfortunately, some of those messages might be lost, messengers could be traitors and generals might either be traitors themselves or honestly disagree on the best plan of action.
"Byzantine fault tolerance" is how well those generals can agree on the same plan despite an unknown degree of potential treachery, disagreement and other misadventure.
At its heart, blockchain is literally just a Byzantine fault tolerance system. It's a way for a bunch of separate entities to agree on the same set of facts. The base blockchain protocols such as Bitcoin, Ethereum, XRP Ledger, Cardano, EOS, Monero and so on are largely differentiated by how exactly their generals reach an agreement.
When a company is trying to choose a blockchain, they're going to look at a variety of factors, including the following:
- Who can be a general and how many generals are there? Ideally, you'll want the freedom for anyone to be a general and as many generals as possible because this would be the highest level of "decentralisation" possible.
- How many messengers the generals can send in a certain span of time. In other words, what's the overall network throughput?
- How resistant the generals' decision-making process is to outside interference and general mucking up. Is the system sufficiently secure for the desired applications?
- How quickly the generals can be near-100% sure that they're all in agreement. How long does it take to confirm transactions with a suitable degree of certainty?
These factors can be adjusted by changing the way generals send messages and make decisions, and the overarching goal of any blockchain protocol is to find a decision-making process that achieves the best possible result across all four of these criteria.
When someone talks about cryptocurrency mining rewards or incentive mechanisms, what they're actually talking about is a system that incentivises the generals to follow whatever decision-making rules the system wants the generals to follow.
One of the reasons it's so tricky is because it's all interconnected. More generals can mean slower decisions or different security assumptions, and some decision-making functions can't be clearly encouraged through the use of incentives. One of the reasons the world hasn't yet seen any serious blockchain adoption is because there are few blockchains that genuinely manage to deliver great results across all of these criteria.
Through this lens, it's easy to see how thoroughly unsatisfactory many of the best-known blockchains currently are.
Bitcoin and EOS generals
In Bitcoin, for example, the generals can send about fvie messengers per second on a good day (five transactions per second). They will end up at whatever decision at least 51% of the generals agree on (resisting 51% attacks), and they can reach an agreement every 10 minutes on average (10-minute average block time) with each decision reinforcing their confidence in the previous choice.
It works this way because Bitcoin's generals make decisions by, basically, having a shouting match every 10 minutes to determine who gets to make the decision each time. They all shout their decision at the same time, and whichever choice is the loudest is what they go for. The volume of each individual general's voice is determined by how much Bitcoin mining power they control. Anyone can come along and join the shouting match and make any decision they want, but whatever decision they make has to reference the previous general's choice.
One of the only sure ways for a traitor to skew the decision-making process here is by shouting louder than all the honest generals combined, across several consecutive decision-making sessions.
Bitcoin is very robust, but as a blockchain, it's completely unsuitable for business use because it's so slow, and you can't make a decision without having a protracted shouting match first.
By contrast, EOS has a system of 30 democratically elected generals, who are elected mostly by themselves. Its system can rush through a lot of decisions very quickly, but you can only be sure that its decisions are correct if you assume that the majority of generals are honest and independent, despite the fact that the election system itself is directly geared towards the creation of cartels.
EOS is fast, but it's completely unsuitable for business use because no one will ever entrust critical business operations to the assumption that a small handful of unknown people somewhere in the world will be completely honest.
The latest and greatest generals
Bitcoin set the benchmark, and from there a lot of other blockchains basically cheated by doing a similar thing with less decentralisation or more lax security assumptions. To make real progress, you need to get creative with how your generals communicate and what constitutes a decision.
Fortunately, there are plenty of ways to do this.
Some blockchain generals have shouting matches while others elect speakers in rotation. Some gossip among each other freely, while others reach a consensus within smaller groups and then have each group cast a vote in the larger group. Some are able to communicate all kinds of peripheral information while others only pass along the essential nuggets of wisdom.
The possibilities are endless, you can combine different developments, and there are dozens of these new systems out there.
- RepuCoin. Coming out of Monash University, RepuCoin achieves 10,000 TPS with tight security assumptions by using a built-in reputation system to restrict decision-making powers to a reliable and proven subset of generals who also meet certain shouting volume thresholds, so any dishonest generals would need to both shout the loudest as well as have an excellent reputation.
- Red Belly Blockchain. Straight from CSIRO, the Red Belly Blockchain achieves 30,000 TPS through the use of a rotating coordinator general who speeds up decisions by suggesting answers for other generals to accept or reject, and calling a decision finalised as soon as it's agreed on by a sufficient subset of generals.
- Fantom. Hailing from South Korea's Yonsei University, Fantom can hit 300,000 TPS by having generals simply accept or reject individual decisions as they pass through, rather than going through a big song and dance around distinct decision-making sessions. Separate subsets of generals can agree on specific decisions in different places at different times.
And then you have Cypherium, which uses a hybrid proof-of-work system called HotStuff. Although Cypherium adopted it first, HotStuff is also the same underlying decision-making system that Facebook settled on for its highly-touted Libra blockchain, with some modifications. So it's reasonable to assume that HotStuff is indeed hot stuff.
HotStuff was invented by researchers at the Swiss Federal Institute of Technology, or Ecole polytechnique fédérale de Lausanne (EPFL) if you're feeling continental.
It can be envisioned as a decision-making system where a leader is randomly chosen from among the generals every time a decision needs to be made. This leader then needs to race around collecting signatures for that decision from other generals. Once they have the required amount of signatures, they rush to a big stack of paper in the middle of all the generals and throw their petition on the top of the stack. There are constantly new leaders being chosen and racing back and forth between the pile of papers and the other generals.
The process is aided by two robots. One of them assigns decisions to generals and hands out the petitions for them to get signed, while the other is responsible for making sure that big pile of papers remains neat and organised even with a bunch of leaders constantly throwing signed petitions on top of the pile.
In only slightly less-fanciful terms, the system works by choosing leaders among the generals semi-randomly whenever a decision needs to be made. These leaders are then responsible for proposing decisions and collecting votes in favour of their decision from a bunch of other generals.
When enough other generals vote in favour of the leader's decision, which they do on the grounds that they've looked at it and said "yep, that's what I would do too", the decision is turned into something called a "quorum certificate" (QC), which basically means all those other generals have approved the decision.
These QCs are then added to the top of a stack of digital papers to be added to the blockchain. Any time a leader needs to put through a new decision, they'll reference whatever previous QC they think is at the top of the stack to show it's a consistent part of the blockchain. If they don't agree with the leader's decision, they won't accept it and the proposal will simply time out and a new leader will be chosen.
HotStuff needs a couple of interconnected mechanisms among its generals to keep this system ticking along as desired.
One is a set of rules for periodically shuffling the constantly growing stack of QCs into a logical order and consolidating accidental side stacks into the main pile. This is because the system is quite frenetic, with leaders constantly throwing new QCs onto the pile. Once a decision has been formalised as a QC and shuffled into that stack, it's officially finalised and part of the blockchain. The reason these rules can work is because you have a pretty reliable set of rules for how frequently a new leader is elected and how long it takes for generals to sign off on a decision.
This ties into another mechanism called the "pacemaker". The pacemaker is responsible for choosing new leaders in a random (that is to say unpredictable) way, and then impatiently tapping its watch while the leader races around collecting signatures for its QC and keeping everyone on schedule.
The point of the Cypherium token in this system is twofold:
- It's given to well-behaved leaders and generals as a reward for processing decisions in the precise way required for this system to work.
- It's required to pay network fees to deter spam attacks. Cypherium users are the source of all the decisions the generals will need to make, so by requiring people to pay fees, you can't deliberately overburden the network by constantly demanding hundreds of millions of simultaneous decisions.
What do people look for in HotStuff and Cypherium?
HotStuff is the name of this decision-making process, while Cypherium is the name of the first company to incorporate the HotStuff decision-making process into its blockchain product.
The end result is that Cypherium is offering a blockchain product that boasts high speed, strong security measures, a high level of decentralisation and a stamp of approval of sorts from both Facebook via Libra and Google Cloud.
Of course, the Cypherium blockchain itself will exist regardless of whether Google Cloud customers are able to access it, and those leaders and generals will keep racing around and doing their thing either way. That's kind of the point.
So the actual nugget of value you find in Cypherium's collaboration with Google Cloud is that Google Cloud customers will have much easier access to this particular blockchain.
It's a sign that Cypherium has probably been vetted to a certain extent, so end users don't necessarily have to crunch the numbers on how exactly the generals make decisions and can instead reasonably assume that it all adds up. And it makes it easier for businesses to build proofs of concept and production blockchain applications that will work on Cypherium without needing to start from scratch.
Plus, wider availability of next-gen blockchains will go a long way towards quenching that persistent "blockchain can't scale" myth, and the proven application of each new generation of blockchain helps lay the groundwork for future improvements.
Most of all, it's easy access to this particular decentralised, automated decision-making machine, and all the incredible new things you can do by having a decentralised, automatic decision-making machine close at hand.
"We're delighted to be collaborating with Google Cloud to provide enterprises with a full-stack solution to harness the potential of this paradigm-shifting technology," said Cypherium CEO Sky Guo. "The growing demand in the market for DLT solutions in the financial industry and beyond drives our commitment to this collaboration. Cloud customers can rest assured that the blockchain solutions they implement using Cypherium Enterprise are clad in robust security, and capable of delivering rapid transaction speeds for its smart contracts and achieving fast data processing from its Java virtual machine."
"Google Cloud and Cypherium are bound by a perpetual need to innovate," Guo added. "The future of commerce and blockchain are inextricably linked and we are well-positioned to leverage Google Cloud's expansive resources and best-in-class infrastructure to accelerate the use of the technology to solve real-world problems faced by businesses today."
Disclosure: The author holds BNB and BTC at the time of writing.