Progressive Decentralization of the RPC Layer: Centralization to Federation

This blog is adapted from Tom Hay's talk at EthPortland 2023, titled "Progressive Decentralization of the RPC Layer: From Centralization to Federation," given on August 18, 2023.

In creating the decentralized infrastructure network (DIN), the first technical challenge Infura, Rivet.cloud, and the rest of our partners have endeavored to do is to transition from various centralized remote procedure call (RPC) services, to a federated network of RPC services. Moving from a centralized architecture to a federated one is how we will progressively decentralize the RPC layer. A federated approach can be defined as:

1. Multiple independent service providers using a routing infrastructure specification to best serve responses.
2. Multiple independent service providers participating in federated governance to bootstrap the network.

A deep dive into the decentralized infrastructure network

Since we are in Portland, I’m going to use the analogies of coffee, beer, and wine to illustrate both the technological and business operations that are required to federate and achieve progressive decentralization. Hopefully, everyone here has had an opportunity at this point to sample one of these fabulous offerings of the Pacific Northwest or is making plans to.

As a blockchain infrastructure provider and as a group of node providers with a mission to create a new decentralized network, we have witnessed that the RPC layer and web3 are rapidly commoditizing. Standardization is also occurring across blockchain RPC providers, meaning that there’s an opportunity for us to come together and define these standards together in order to achieve a decentralized RPC protocol that's not only resilient and performant, but also economically viable, for all.

Coffee itself is a commodity. However, there is a distinction between craft or specialty coffee and commodity coffee. When thought of as a commodity, this is the coffee that is bought and sold on exchanges. It is often thought of in terms of coffee as a product, and is interchangeable - there is not much thought given to the specific lot where that coffee is grown.

If you have spent any time here in Portland, you know that we love our specialty coffee. This is coffee where the origin of the coffee is known; it is likely purchased from a specific microlot sold by the farmer to a specific roaster. There is a specific flavor profile associated with it, and people here are really excited to try coffees from specific regions, growers, and meters above sea level.

The processing of the coffee - whether it is washed, honeyed, naturally processed, etc., is also a factor in the taste, as well as the varietal of the coffee. We love to drink specialty coffee here and see great value in it. In fact, we are willing to pay more than the commoditized price of coffee for that specialty.

We also love our craft coffee. Craft coffee is specially prepared coffee, made by some of the best and most talented coffee professionals in the world. Craft coffee includes any of the drinks that require knowledge and skill to make, and are made to order.

Coffee is the clearest example of this, but to a lesser extent, there are similar dynamics at play in the world of beer and wine. In beer, there is the prevalence of craft beer - which started as an alternative to macro-brewed beer that, while not truly a commodity, there has always been a back and forth between the craft/specialty beer and more mass-produced options.

Wine has more so defined itself through the specialty distinction, putting a lot of emphasis on the terroir, the year of production, etc. But there are also blends of wine that are priced cheaper and that are canned for accessibility.

We believe that there is a clear space for the commoditized RPC layer - where RPC providers would compete directly on price for a certain level of quality and service. Certain consumers would look for what price meets their expectations for the level of quality they seek. We think a decentralized approach will raise the level of quality available to consumers while keeping prices fair.

We also believe that there is an opportunity for specialty or craft RPC. This could be the geographic component - which becomes important when there are considerations regarding latency, jurisdiction, and provenance. This could be higher-level features offered - such as indexed logs that are available for certain high-traffic networks. This could also be the availability of networks themselves - new protocols that may only have a public RPC hosted by the protocol but are looking to grow the availability of reliable RPC providers.

Federation first, then decentralization

We are going to work towards a model of decentralization, but our first step is going to be federation.

This is where competition and collaboration become key - and I’m going to return to the example of coffee and the community of coffee roasters within Portland. We have a few facilities here that are called co-roasting facilities - like the Buckman Coffee Factory and Roast, where multiple coffee companies roast their beans in shared facilities. They are able to rent and share equipment in order to defray costs, or focus on what they are best at, which might be sourcing great coffee from certain small farmers using interesting methods that produce unique beans.

In brewing and in wine, the same is also true - there is a fair amount of either contract brewing or contract winemaking - where a brewer will come to a larger brewery with a recipe and pay them to make it and can that beer, or a winery will rent excess space at a winery to produce wine for a season - there are multiple examples here. In beer, there are plenty of collaborations as well, like Breakside Brewery, which makes a beer called WanderJack with a brewpub in east Oregon called Barley Brown’s.

I use these examples because what we are doing with the decentralized infrastructure network is analogous. We are working collaboratively, and also somewhat competitively, with a team of infrastructure providers that each offer different feature sets to their users. We will compete with each other to produce the best services among ourselves and will also work together to offer our customers the greatest number of networks, with the most uptime, and the most opportunity to get access to “specialty” or “craft” options.

And this first stage is going to follow a similar process to what occurs in all three industries currently - a process of progressive decentralization through federation. We will route requests based on a federated system of providers with distinct offerings. We will be offering new options to our customers based on what our federation is able to offer, including specific geographies, and new chains.

The task at hand will be to prove that consumers would pay for these services, and obtain value from them. As we do that, we will concurrently be taking the steps to build a decentralized system in which any provider can participate, even those that already have a decentralized model. We will also work on the software supply chain in order to ensure that users are safer than ever.

I hope you get a chance to sample our coffee and other beverages while you are in town (if you don’t consume caffeine or alcohol, almost all of our places offer alternatives that are still very tasty and refreshing). This is how I think about the decentralized infrastructure network, and I hope you continue to follow us on our journey.