In this piece, we’re going to dive into the overarching vision fueling Constellation, while using our CTO Wyatt Meldman-Floch and our V.P of Engineering Ryle Goehausen’s rousing talk from TechCrunch as source material. We’ll address the current limitations existing in the distributed ledger space, and how our approach differs. If you haven’t had a chance to check out their presentation from the event, give it a watch below.
First off, what are the current limitations?
Let’s start at the beginning with a quick breakdown of DAG technology and a recap of the origins of blockchain technology. Blockchain technology, as it currently stands, allows you to secure digital transactions on a ledger (Bitcoin) and then moved into securing digital contracts (ETH). Now, we’re finally trying to solve the problem of applications, which not many of the existing technologies can do. One of the main issues with existing blockchain infrastructure is that they just don’t work at scale, and aren’t ready to handle the massive scaling demands at the enterprise level. The issue of scalability is another that we’re strongly focused on solving.
Alongside the likes of HyperLedger (who we recently partnered up with), InterLedger, Cosmos, and Parity, Constellation shares a similar complaint regarding how Ethereum is skewing more towards an excessive amount of centrality. “Plenty of people have talked about Ethereum moving towards too much centrality, and this is probably due in part with how much money and attention they have raised, they’re trying to push the bar past what Bitcoin can actually offer,” said Ryle.
Ethereum and other protocols lock you into their “centralized sandbox,” where different networks cannot connect, and you’re forced into using their coding language. Some other common blockchain limitations are resource waste, redundant work, and running application code over and over. Ryle went on to share how, as it currently stands, “you can’t really host a modern application on Ethereum. You can serve a website, but you couldn’t really serve a scalable database. It wouldn’t work. These networks can’t connect. You’re locked into that abstraction… there is no throughput to speak of.”
Enterprises already expect this level of functionality and fluidity in the computing infrastructure space, and it would make sense for them to expect to see a similar shift happen in the blockchain space before diving in head first. “Everybody is coming into [blockchain] expecting it to be like it developing on AWS, and it’s just not – it doesn’t work that way right now,” said Ryle.
“You can’t have competing VMs. You can’t have competing execution providers. You can’t deal with people who don’t want to run code from other people, as every node needs to run all of the applications,” said Ryle. In addition, what incentive do you really have to build an application on one of these protocols, when all you’re ultimately doing is enriching the creators of that protocol?
“This is unfair, and it avoids all of the issues of integrating properly and building a community if all of the value is driven into the protocol.”
Ryle went on to mention how with double spend attacks, “any miner who has enough power can instantly destroy” a smaller network. “Anyone with enough stake can manipulate their markets and the consensus models don’t reinforce one another.” These types of consensus models are also highly anti-entrepreneurial. “You can’t start your own network. You have to go to a large provider. If you have a new idea, you have to do what they say. You can’t just create a network and try and integrate it with something, because they don’t trade resources fairly. There is nothing that allows these networks to trade security with one another, and reinforce each other fairly,” per Ryle.
To reiterate, the issue with these larger protocols is that they have to run all of the applications inside of them. They lack application separations and boundaries between them, which also creates data management concerns. How do we solve this issue? It all comes down to the notion of consensus as a service. “The way to abstract over this is to focus on isolating out consensus in and of itself as a service, and focusing on the performance of that because that happens anyways in the regular ecosystem.”
While Constellation ’s solution is based around the use of DAGs (directed acyclic graphs), we would slightly rephrase this argument as it being nonlinear. “If you’re talking about consensus across networks, there are not necessarily linear relationships. It’s coarse, it’s granular. It’s graph-like in nature, and they’re globally distributed networks. Each application has its own unique access pattern for data dependencies and all these other issues. The huge thing here that doesn’t get talked about is if you create a small network, it’s not going to be secure.”
“Consensus as a service” underpins Constellation’s approach. As Ryle puts it, “there should be a proper boundary created so that you could use varying VM’S or contract languages that might be suited to different platforms — like running on a cell phone, running in the cloud, running on a home computer — these are all different heterogeneous environments.”
Ryle went on to add some further context as to how Constellation is taking a unique “open-sandbox” approach, especially in comparison to Ethereum:
“In simplest terms possible, Ethereum does not support forks of itself. We want to have people re-use our code to build their application-specific networks and then integrate with ours, which is a different model. Under Ethereum, they set the rules for you to play in a sandbox. You don’t get to mess with Ethereum’s core code. With us, we allow you to extend your application as a ‘custom version of Ethereum’ and provide a mechanism for integrating it with our core network.”
While a number of other protocols have taken on so much ground and make such lofty claims that they can’t solve anything practical at all, we want to avoid that trap altogether. Rather than promising to “revolutionize and disrupt” all the worlds issues, we’d rather focus on solving consensus at scale.
As Ryle succinctly puts it, “if you can solve consensus at scale, there are any number of plug-and-play components that you can incorporate from other projects which is far more interesting from a decentralization perspective. We’re interested in networks reinforcing one another rather than dictating what you have to do.”
When you pare down what blockchain is, it’s a way of securing updates to a global database. When it comes to enterprise-grade adoption, the space as a whole isn’t ready for that leap, but in the meantime, we can look towards AWS as an ideal role model in terms of having an enterprise-grade security standard. Ryle went on to say how AWS’s security model doesn’t enforce security to the degree that it forces you into only using specific solutions, but rather, it’s designed to empower all sorts of different solutions. “AWS does not have a completely unified platform where you have to do everything a particular way – they offer plugins and customizations to developers”.
Ryle went on to share how he’s noticed a considerable problem in platformization, and by that he means how there is an underlying centralization trend among decentralized networks, saying “as they become larger, they become more centralized – this is crazy.” In our view, it’s more important to focus on disparate networks and to integrate different types of networks and microservices as an approach. Also, when more nodes and actors join our network, we become more decentralized as opposed to more centralized, simply by the inherent nature of our DAG infrastructure and our reputation based consensus model.
PoW vs. PoS vs. Constellation’s reputation model
Lastly, let’s examine some of the fundamental differences between the two main consensus models being utilized in blockchain right now – Proof of Work (PoW) and Proof of Stake (PoS). With PoW, the origins of this can be found in anti-spam for email. “It would be the death by a thousand cuts, where if you sent out a large amount of computation – it becomes cost prohibitive to spam a large number of people,” said Ryle. PoW was originally intended only to be used in a small, rate limiting capacity, and double-spend attacks regularly occur on smaller networks. “These double spends are in the news all the time on smaller networks. It doesn’t work. It’s a known vulnerability.”
Proof of Stake, while it addressed the wastefulness issue, it still has vulnerabilities. Smaller networks can still be easily attacked, and in terms of governance, it’s fundamentally anti-democratic. “We don’t run the world by trusting people just because they have money, right? That’s not Democratic. That’s not a governance solution. And this goes beyond just invalidating individual transactions and having penalties. It’s more about the governance of the entire protocol itself. You can’t rely solely on buying votes, that would be considered absurd in the real world,” said Ryle.
Regarding Constellation’s novel reputation based consensus approach, trust and reputation scoring is already conventional throughout society. “This is credit scores. This is auditing, this is accounting, finance – this is considered normal.” Although this notion is prevalent throughout society at large, there has yet to be a strong evolution in this and it hasn’t been codified properly into digital form. Our approach allows for an enormous amount of machine learning that can be done, that enables truly democratic governance. “Who are the people involved in these things? Where are their IPs coming from, all sorts of off- chain data that can be integrated very very quickly and easily – so far that really hasn’t happened,” per Ryle.
As it stands, there is still a massive level of ignorance about what happens in the code publishing process and who actually governs the project itself. “All of these steps require trust and they just they just ignore it completely. We need to go towards Democratic governance, not buying votes,” according to Ryle.
To sum it all up, the current blockchain solutions just aren’t ready to scale, and their underlying consensus models are vulnerable to attacks and are anti-democratic in nature. With Constellation, we’re aiming to create this notion of consensus as a service, while also solving the main issues of scaling for enterprise use and creating a truly democratic governance model. We’re excited to bring this vision to life, so be sure to continue to follow our journey at any of the links below!
Become a Constellation Member
Participate in our Community
Check out our Technology