The BOSI Stack. Blockchain Open System Interoperable Stack

The physical world has many constraints. Each jurisdiction has its own rules, language and currency. Since the inception of the internet in the 90’s, we’ve had amorphous digital businesses and locations whose owners decide the currency, rules of engagement and boundaries within these new digital jurisdictions.

Web3 as an ideology and blockchain as the enabling technology has provided the opportunity to build more formally defined digital jurisdictions above our physical ones, allowing for more creativity and expression in how these jurisdictions are governed. We now have open-source, permissionless platforms that mirror public commons, a digital commons where rules are either changed by the will of its participants or forked off into new ecosystems. These jurisdictions are much more clearly defined, through hard and fast protocol rules. People can freely choose to participate in as many of these digital jurisdictions as they want.

As a result, we have seen an explosion of open-source protocols and are inevitably headed towards a multi-chain, interoperable digital world layered on top of the world of atoms and bits. We are starting to see patterns in which these new ecosystems are organizing themselves. However, before we get into that, we must first we must look into what the Web2 world was built on.

To help understand why these ecosystems are forming the way they are, it is helpful to first consider the OSI model from the Web2 world. I am not the first, nor will I be the last to see the similarities between the blockchain industry and the OSI model, that being said it is helpful to consider the similarities between the two approaches when considering how the Web3 space may evolve.

OSI Model

Developed by the International Standards Organization, the Open System Interconnection (OSI) is a conceptual framework that allows us to think about how the networking stack of the internet is organized. The clearest explanation I could find comes from Vangie Beal:

It divides network communication into seven layers. In this model, layers 1–4 are considered the lower layers, and mostly concern themselves with moving data around. Layers 5–7 called the upper layers, contain application-level data. Networks operate on one basic principle: “pass it on.” Each layer takes care of a very specific job and then passes the data onto the next layer.

The OSI Model describes:

  • The way in which devices communicate with each other.
  • The means used to inform devices when and when not to transmit data
  • The methods which ensure that devices have a correct data flow rate.
  • The means to ensure that data is passed to and received by the intended recipient.
  • The manner in which physical transmission media is arranged and connected.

The OSI model provides a rigid communication standard that describes how the internet works today and can be used as a corollary for the way the crypto industry is shaking itself out. We can *loosely* separate it into hardware-focused protocols where participants are incentivized to use and run hardware; and end-user focused protocols like social media, messaging, gaming and insurance built on top of this physical infrastructure. As the industry matures the communication between each layer will ossify leading to more refined methods of building on top of the Web3 stack.

Blockchain Open System Interoperable stack

I’d like to introduce the Blockchain Open System Interconnection stack (BOSI). The BOSI stack is a loose framework, where decentralized protocols sit on top of each other in a stack. They can implement various features as interchain communication improves, or leverage other sets of protocols in the same blockchain.

The goal of this framework is to answer the following questions:

  • How do independent blockchains and protocols interact with each other?
  • How do developers and product designers leverage interconnected pieces of the stack?
  • Where are there opportunities to build and invest in new protocols?

While important, scaling solutions like plasma and state channels are not included because they act as vehicles to make accessing core protocols easier and more scalable, and are not, in effect, the end solution themselves. Additionally, this is not meant to be a comprehensive list as there are likely dozens of other layers in the stack that are missing that others are building.

Unlike the OSI model, the BOSI model is much less rigid. There are no predefined interfaces that force one layer to interact with each other. Developers can choose which part of the stack they want to leverage. When building a decentralized application, pieces of this stack can be swapped out for traditional, centralized solutions, depending on the need and the maturation of its decentralized counterpart. This will continue to be the norm until there are mature, stable protocols that are user-friendly.

The shared layer has its own set of solutions which are generalized protocols (e.g. Aragon, Colony, dxDAO) can be implemented independently by a specific protocol, like Politeia with Decred. As solutions like reputation, anonymity, identity and governance mature, new protocols will find fewer reasons to reinvent the wheel and use what exists.

For example, Decentraland is an end-user protocol which may leverage The Graph for its caching layer, IPFS for storage and DAI to purchase land, all built trustlessly using Ethereum smart contracts and consensus. Decentraland could use another protocol for identity and Aragon for its governance layer.

Another example, Decred is a payments cryptocurrency that has built its own native governance layer within its protocol called Politeia. As a part of the utility of the protocol they also plan to add a DEX that utilizes on-chain, atomic swaps trustlessly. There are centralized pieces of the Decred stack, such as the block explorer API that developers can use to build wallets and applications. All built using its own unique, hybrid set of Proof of Work and Proof of Stake consensus.

These examples show that protocols can pick and choose pieces of the decentralized stack they want to use, and as the industry matures over time the options will continue improving.




The decentralized stack: A magnitude more diverse than the OSI Model

In all but one layer (application layer), the OSI model has, at most, 2 dozen protocols within an individual layer. As you can see below, all of the protocols under the Session layer:

To map it over to the BOSI stack, each bubble is considered a layer and has at a minimum, one protocol to service demand. The trend over time has demonstrated that as market interest increases, the result is in an explosion in competing flavors within the same layer. We saw this with bitcoin and its thousands of forks, Ethereum and dozens of competing smart contract platforms, and the dozens of stablecoin and lending protocols built on top of Ethereum today.

We are starting to see this trend continue further down the stack, where more hardware-specific incentives, use cases, and protocols are being launched. We see this with projects like Althea, Helium, Livepeer, Pocket and hardware-specific solutions like DAppNode and Coinmine.

Over time, we’re going to see these protocols get more and more competitive as we see more success within individual categories along with more replicable governance and economic models. Like smart contracting platforms, we will potentially see dozens of protocols in each layer of the stack, leading to a magnitude larger set of protocols interacting with one another freely.

Application-specific blockchains (ASB’s) have a big role to play here as well. As it becomes easier to leverage proof of stake consensus algorithms like Tendermint and Grandpa and as economic models and governance models mature, building a blockchain will be like building your own PC. This will particularly be the case with protocols sensitive to the cost of coordination on smart contract platforms (e.g writing state to Ethereum is expensive) and tradeoffs in decentralization.

A key difference is that unlike our real-life jurisdictions, we can participate in any number of these protocols. The question of whether traditional power law characteristics and network effects apply in this new arena remains unanswered. There are definite network effects from financial systems and staked tokens, however robust derivative markets could dramatically change these network effects over time. This leads me to think that protocols will mature more with regional and local network effects that rely on community and belief in the story of the protocol.


We are in a fascinating period of the app=>infrastructure cycle. The industry has not had any real application level hits since Cryptokitties in 2017. Previous applications built using Solidity are moving to ASB’s. When seeing the BOSI stack fill out, we’re seeing a race of ASB’s (Aragon, Centrifuge, Handshake, Witnet, Pocket) against smart contracts on scalable platforms.

Building an ASB opens up the design space and potential for efficiencies — protocol designers don’t have to work within the constraints of the EVM (while still being compatible with it), effectively enabling function calls to be “closer to the metal” in terms of efficiency. This is massive when the costs of coordination are critical to compete against their centralized counterparts.

Much of the inefficiencies we see are on the hardware layer. Blockchain clients are not optimized for scaling and are expensive to run, though we are seeing projects like Rivet apply proven web2 models like streaming replication to improve on this. New economic models through optimized ASB’s or cheaper smart contract platforms will compensate while these inefficiencies are improved upon.

Already, with just the few layers described here, we’re looking at hundreds of existing protocols solving specific problems. We’ve seen the DeFi space explode with different protocols for the same layer due to where we are in the evolution of scaling for Ethereum. We will see the same explosions up and down the stack as we have more successful governance and economic experiments within the space.

I have never been more optimistic about the space from the conversations I’ve had with projects creating novel designs. P2P networks combined with token models have the potential to reduce cost structures by an order of magnitude that centralized companies cannot compete with.