research

The Convergence Stack

March 2019

Posted by

Lawrence Lundy

Partner / Head of Research

He helps inform the UK and EU on technology policy as an expert advisor to the UK All Party Parliamentary Group (APPG) on Blockchain; steering committee member for the AI Global Governance Commission; community member of the UK All Party Parliamentary Group (APPG) on AI; member of the European Commission Observatory on Blockchain; and contributor to the UK Cryptoasset Taskforce led by HM Treasury, FCA and Bank of England....read more
research

The Convergence Stack

Posted by Lawrence Lundy - March 2019

 

[stack] Noun. an ordered pile or heap

The Convergence Stack is a further refinement of the Outlier Ventures investment thesis first published in 2016 anticipating the convergence of blockchains with IoT, AI and eventually robotics. Over the last three years, as the industry has matured, we have developed a more granular view of the specific technologies and protocols that make up Convergence and that informs how we invest.

 

As a direct consequence of those investments and our work with startups, a growing corporate ecosystem, and leading academia, we have come to think of our portfolio in terms of a literal stack of highly synergistic technologies now framed as ‘The Convergence Stack’. A set of interlinked and open-source technologies spanning hardware, software, networking and applications that support a more secure, private, accessible and ultimately what we hope to be equitable digital infrastructure.

 

However, the task of evolving and simultaneously realising such a vision is simply too colossal to be done in isolation. Instead this post serves as an open invite to engage both with us and the stack of technologies, their founders and communities to extend, deepen and diffuse their innovations. We hope the stack will be realized in concert with a diverse set of stakeholders from academia as well as early and later stage startups, to the largest corporates, supported by forward thinking regulators and governments.

 

Because this is an open stack anyone can contribute to its codebase and because at their core these are decentralised networks anyone can contribute to their growth and subsequently participate in their economic success.   

The Convergence Stack Outlier Ventures

 

Stack Explorer

This high level design of the Convergence Stack is a starting point. We went through several iterations creating it, and imagine there will be many more possible versions as we collectively build it out together with our portfolio teams, especially in such a rapidly evolving technical and socio-economic environment. Our objective isn’t to be definitive but rather to add to the debate.

 

We have used the familiar OSI model to integrate these new protocols and networks in the broader technology industry rather than something separate and parallel. Although we made some tweaks to the layers because it doesn’t map perfectly. Nevertheless the stack follows the same general structure from the physical layer at the bottom to the application layer at the top. Many projects will fall into multiple categories as they use combinations of technologies to solve their customers’ problems. As the market matures and consolidates, projects will look to vertically and horizontally integrate and aim to have multiple products across the stack.

 

Whilst the stack is open to anyone to develop and inform we will be looking to drive forward its evolution through a number of core working groups under the banner of The Convergence Alliance, made up of the respective protocol CTOs, integrators and enterprise customers which will be announced in some follow up posts by Outlier Ventures’ CTO Aron van Ammers.

 

Hardware Processing, Hardware Storage and Hardware Networking

At the bottom layer, it’s likely that processing, storage and networking will all be integrated into an IoT device or machine. For clarity, we have chosen to break  them out as separate components. First, secure hardware processing uses trusted execution environments (TEEs) to encapsulate confidential computations. We are seeing increased use of hardware security modules (HSMs) and trusted execution environments (TEEs) such as Intel SGX by the likes of Oasis Labs with Ekiden and Enigma. We expect to see further innovation in making TEEs truly secure as well as moves towards open-source chip architectures like RISC-V. Secondly, hardware storage products like Trezor, KeepKey and Ledger are already used to store private keys, and as the market matures it’s likely hardware storage will ultimately end up embedded into mobile devices and wearables. The Samsung S10 and the HTC Exodus are already leading the way in this regard. Thirdly, an often ignored point of centralized control is hardware networking. No matter how secure we make software, bits and bytes will need to move through a physical network of gateways, switches and bridges. Projects like NetFPGA, the Open Compute Project, and Telecom Infra Project (TIP) are pushing secure open-source designs in this area.

 

Ledgers, Storage and Networking

Above the hardware come ledgers, storage and networking. Depending on functionality ledgers can be blockchains, smart contract platforms or dApp platforms. This is the area that seeded the broader ‘web 3.0’ movement and the source of fiercest competition today. Projects like Ethereum, Dfinity, Hashgraph, Sovrin and Fetch.AI are differentiating on scalability, security and decentralization; making different trade-offs depending on the targeted applications. Some are prioritizing scalability like IOTA and Zilliqa, others security like Tezos and Cardano, and realistically only Bitcoin is committed above all else to maximising decentralization. There are of course private chains like Corda, Quorum and other blockchain-as-a-service (BaaS) solutions focusing on enterprise customer requirements where with known participants decentralization is less of a concern than security and scalability.

 

Storage is a well-researched area in decentralised systems. IPFS is the current go-to provider, but there are many tokenized solutions like Siacoin, Storj and Maidsafe as well as the upcoming releases of Filecoin and Swarm. Moving right, the final technology at this layer is networking. Software-Defined Networking (SDN) is typically ignored in the Web 3.0 space, but projects like OpenFlow and Facebook Open Switching System (FBOSS) are examples of widely used open source projects that manage networking equipment. Open-source and ideally community-driven SDN is vital in ensuring there are no centralized chokepoints in the overall stack.  

 

Scaling, Databases and Bridges

As we move up a layer we come to scaling, databases and bridges. Firstly, we are considering any project aiming to increase the performance of the underlying ledger as ‘scaling,’ which includes Lightning, Plasma, and bloXroute. For simplicity’s sake we consider any type of state channel and sidechain to be included here. Secondly, there is an important role in the stack for decentralized databases like OrbitDB which ensure structured data is stored and shared in an accessible, peer to peer way. Finally, bridges like Interledger, Cosmos, Polkadot, and AION ensure value and data can move across ledgers. Ultimately technologies at this layer are focused on making the decentralized web faster and interoperable.

 

Authentication, Query and Compute

Next up is authentication, query and compute. Authentication can be considered the identity layer, and includes namespace registries like Handshake along with attestation like Evernym, uPort, and WebID. It’s likely we will see a Facebook identity service enter the fray at some point, and Microsoft has been an active player in the decentralized identity ecosystem over the past few years already. Discovery and curation will become critical as the amount of data in decentralized systems increase, so what we call Query includes TheGraph as well as data verification tools and oracles such as ChainLink, XYO Network and FOAM – essentially any tool that enables the indexing, querying and verification of decentralized data. Last, most computation on data must be done off-chain. Confidential and verifiable compute includes zero-knowledge and Multi-Party Computation (MPC) implementations like TrueBit, Enigma and Starkware. Over time we expect a tighter hardware and software integration to ensure end-to-end security.  

 

UX, API and Middleware

Sitting above the lower level tools but below the application layer are UX, API, and middleware. UX is how users engage with the decentralized web; we consider projects like Brave, Metamask, and Status to be UX tools. Browsers, wallets, Conversational User Interfaces (CUIs) and spatial computing will all play a crucial role in abstracting away complexity and onboarding new users. APIs and other back end tooling like infura, DAppNode, and VIPnode are vital in ensuring the Web 3 space is accessible to the full community of open-source developers. Projects like Codius attempting to provide an open-source hosting platform also fit here. Finally, also sitting on this layer is middleware, platforms for building domain-specific applications. Examples include Dharma and MakerDAO in finance, Aragon in governance and Augur in forecasting.

 

Applications, Marketplaces and Learning

Finally, the top layer includes applications, marketplaces, and learning. Firstly, applications include all of the dApps as early adopters we know and love including the NFTs (non-fungible tokens) services like CryptoKitties, applications using middleware tools like Veil and Guesser, productivity apps like Graphite and financial services like Balance.io. Secondly, Marketplaces like Ocean Protocol, Streamr, Wibson and Tide Protocol fit here, ensuring data liquidity for the rest of the stack. Finally, machine learning is a core component of the Convergence stack. We have always believed that just like in Web 2.0, one of the most valuable places in the value chain is making use of the data available at lower layers of the stack. We expect real value to be created across all three machine learning paradigms – supervised, unsupervised and reinforcement learning. Projects like OpenMined and others are beginning to combine learning tools with decentralized storage and private compute to build more open, accessible and decentralised AI systems.

 

Get in Touch

The Convergence Stack provides an open-source hardware, software, networking and application stack to support a more secure, private, accessible and equitable digital infrastructure. We want to partner with, support and invest in projects working in any of the areas including early stage startups, later stage projects, and corporate projects. If you think you are working on critical infrastructure that will accelerate this stack and vision we want to hear from you.

  • Investment enquires: investment@outlierventures.io
  • Research enquiries: research@outlierventures.io
  • Enterprise and partnership enquiries: partnership@outlierventures.io
  • Marketing or event enquiries: marketing@outlierventures.io

To listen to a full explanation of the stack LIVE we invite you to a virtual conversation with two of the core components: Toby Simpson, CTO of Fetch.AI, decentralised AI for the machine economy and Ken Ebert, Software Architect and OS Engineer, Sovrin the new standard for digital identity. They will be showcasing a LIVE demo of the highly synergistic technologies in practice, introducing their latest project: ANVIL. Sign up: https://outlierventures.zoom.us/