Subspace is an open, scalable, and interoperable layer-1 blockchain platform for storage and computation. Addressing the blockchain trilemma, Subspace utilizes a Proof-Archival-Storage (PoAS) consensus mechanism to provide decentralized storage for protocols and application developers. Node farmers on the network collectively store replicas of the blockchain history, operating as a single distributive archival node. Subspace Network is reimagining Web3 smart storage at the internet scale.
Weilei Yu, CMO at Subspace Labs, joined us for an AMA on September 27th.
vVv: Would you mind giving us a brief overview of Subspace and your role on the team?
Wei: My name is Weilei Yu, and I’m currently the Chief Marketing Officier (CMO) at Subspace Labs leading community growth, consistent global growth, and other go-to-market and marketing initiatives.
Prior to Subspace, I led growth at Dapper Labs. Many of you may recognize CryptoKitties, the creators of the NFT standard ERC-721. ERC-721 served as the backbone of NFT growth and success over the last 2 years. Dapper Labs also launched the Flow blockchain, NBA Top Shot, NFL All Day, UFC Strike, and many other global IPs. At Dapper Labs, I led Flow Blockchain’s consistent growth from pre-launch to nearly 5 million Mainnet users with the deployment of 500 dApps on the blockchain.
Flow became a globally distributed ecosystem with multi-billion dollar project valuations built on top of it and many DevOp activities. Before Flow in 2017, I worked on another crypto project built on Ethereum. Thus, I’ve seen a little evolution across the living ecosystems and the rise of alternative layer-1s over the years.
At a very high level, Subspace Network is a highly scalable infrastructure layer for Web3 to empower and connect other blockchain ecosystems. That is probably our most significant differentiator. We are not here to compete with different ecosystems; our primary goal is to enable and connect ecosystems that add value to us.
Some of the core use cases include data availability for rollups on Ethereum. Rollups are also known as layer-2s, and from what we’ve seen so far, they will likely be the scaling solution for Ethereum. One of our use cases is archival storage for all layer-1s, meaning Subspace can store the entire history from hundreds, if not thousands, of blockchains.
We have already integrated with the Polkadot and Kusama blockchains and storing of their history. Next, we want to integrate with Ethereum. And thirdly, the most critical service we will roll out in the future will be a trustless cross-chain bridging solution to enable our true multi-chain future. Building a robust trustless bridge is very difficult, given all the challenges and hacks that have happened recently. This is a very high-level overview of subsets. As we integrate with other blockchain ecosystems, the applications built on top can easily plug into the Subspace Network to connect directly to different ecosystems without migrating to another chain, hub, or alternative.
”Subspace Network is a highly scalable infrastructure layer for Web3 to empower and connect other blockchain ecosystems.
The most critical service we will roll out in the future will be a trustless cross-chain bridging solution to enable our true multi-chain future
vVv: Can you give us an overview of the current team size and plans for expanding and growing your team?
Wei: Currently, we are close to 30 people. We are onboarding new members weekly or finding talented candidates. The plan is to double our team size by next year. We have ambitious growth goals but are cautious because growth is not beneficial
vVv: What’s the current state of your project? Can you give us some insights if you’re still on Testnet or Mainnet and how many people are currently testing or participating?
Wei: We are currently on the second phase of the incentivized test level, which we call testnet Gemini II. We finished Gemini I back at the beginning of July, where we had over 13,000 nodes run by the community, covering 50+ countries worldwide. We have heavily battle-tested our network unprecedentedly because we want to prove that we are the most secure distributed data availability solution provider of generalized cross-chain messaging.
With Gemini II, we have already seen 25,000 plus nodes run by the community from around the world just one-week post-launch. After Gemini II, we will have a third phase called Gemini III. In phase III, we aim to test everything we plan to deploy on the Mainnet. We don’t have a date for Mainnet deployment, but we’re aiming for Q2/Q3 of 2023.
vVv: How is the testing incentivized? What do people get for running those nodes?
Wei: During Gemini I, we wanted to test every possible scenario under the hood, from Raspberry Pi to data center size, storage implementation to different levels of connectivity, other regions, firewalls, etc. We learned a lot from Gemini I, and with Gemini II, we wanted to focus on the consensus to see how massive of a network we could bootstrap with our consensus model and the farming node we have developed so far.
Instead of being a PvP model, where everybody competes to farm and earn as many rewards as possible, Gemini II is a collaborative process. We have broken Gemini II into three milestones based on the number of nodes, storage pledged, and how many nodes are live at any given time, known as uptime.
There’ll be different tiers of milestones being unlocked, with the last one unlocking at 75,000 nodes and giving every participant the maximum rewards possible. In that aspect, the reward is split evenly across all the participants, and every higher milestone unlocked will give each participant more rewards than the previous one. So ideally, that builds a much more collaborative and friendly environment for people to join.
vVv: Do you intend to have Bitcoin miners provide you with nodes eventually?
Wei: I would like to preface this by saying two things. We refer to node operators as Farmers because of the Proof-of-Capacity (PoC) model. Secondly, the Subspace Network Orion has been permissionless from day one to incentivize testing, which means anyone can join. We have already received a lot of interest from miners with petabytes of storage capacity and mining power.
However, we believe it’s most important to prioritize smaller node operators first to ensure large distribution and decentralization on both the network and participation by future token holders. We feel that it is inevitable that BTC farmers, large and small, will provide nodes on Subspace.
”We believe it's most important to prioritize smaller node operators first to ensure large distribution and decentralization on both the network and participation by future token holders.
vVv: Do you have competitors out there or companies trying to achieve the same goal?
Wei: There are competitors across each use case I mentioned, but we’d love to work with them to expand that in a modular blockchain paradigm. Considering everything we aggregate, you won’t find a direct competitor because Subspace is a modular layer-1 blockchain that can unlock massive horizontal scaling because of the decoupled execution. We want to archive the entire history of each layer-1 because running a full node for different chains gives you an incentive reward. Still, no one is rewarded for running archival nodes because they’re massive and lack additional incentivization.
”Subspace is a modular layer-1 blockchain that can unlock massive horizontal scaling because of the decoupled execution.
vVv: How do you plan on attracting developers?
Wei: Developer adoption for any layer-1 is always the most challenging part, and increasing community participation awareness is essential. However, we feel getting people to commit resources and time to build on top of your chain is quite challenging, which is where we differ from other layer-1s. Our initial go-to-market strategy is less about convincing people to build on top of Subspace and more about Subspace integrating with other layer-1 blockchains and providing the dApp support. Our target audience is not just developers willing to be on top of Subspace but anyone building on the supported ecosystem. We’re prioritizing political integration so we can eventually support hundreds and thousands of dApps.
”Our initial go-to-market strategy is less about convincing people to build on top of Subspace and more about Subspace integrating with other layer-1 blockchains and providing the dApp support.
vVv: Is it more important to focus on having enough nodes run by enough different people for everything to be scalable and to be able to run all those different depths?
Wei: Correct. You need to be trustworthy, reliable, and secure enough. We must ensure that Subspace is robust in one of the most decentralized and distributed networks. In addition to providing a trust-minimized or trustless bridging solution back to Ethereum or other chains, these components are the minimal requirement for some of the big players to integrate, and that’s why we’re working so hard in bootstrapping their network layer first instead of rushing and doing everything else.
”Providing a trust-minimized or trustless bridging solution back to Ethereum or other chains [is a] requirement for some of the big players to integrate, and that's why we're working so hard in bootstrapping their network layer
vVv: We’d like to delve into your funding state and the type of partners you have. What kind of partners are you looking for? What is your overall runway? How do you intend to have enough liquidity to keep building Subspace?
Wei: Subspace raised over 35 million in Q1, led by Pantera Capital. This gives our current team size over a decade of runway. And even as we continue to expand on our ambitious goal of doubling the team size by next year, we still have many years left of runway. That’s not without considering additional funding or a community token offering that might happen before Mainnet. We’re very well-positioned to continue developing and delivering on the roadmap. We’ve had many investors reaching out to us interested in learning about the next fundraising round. However, we have prioritized internal milestones like Gemini II before considering future fundraising.
vVv: What role do tokenomics play in what you are currently building? There’s always a delicate balance between the tokenomics for the ecosystem, investors, stakeholders, and the tokens for the team. How do you currently see those dynamics playing out in the future?
Wei: We believe that we have a function of value distribution. As far as participation in the long term, it’s a great way to coordinate a highly distributed community ecosystem. We are also very community-centric, which I like about the Subspace team. We have allocated more than 50% of the total supply to the community and farmers, which means in the long term, farmers in the community will control the network. Then we have an upcoming foundation similar to other ecosystems that will run grants and different ecosystem growth strategies. We have a few layers in terms of pure tokenomics, but I don’t want to release too many details as it’s still a work in progress. However, we can share more information regarding tokenomics in the coming month. Currently, the farmer and the block rewards are similar to many proof-of-work (PoW) models. Identical to the PoC, our PoAS is much more environmentally friendly and sustainable. There’s the block rewards subsidy, gas fees for transactions, and storage fees to reward the farmers. Once we enable the trustless cross-chain messaging protocol, there will be an extra fee for message passing. In that sense, there’ll be a few value captures across different layers.
vVv: Why did you decide to develop on the DOT ecosystem? And what are your plans to become cross-chain?
Wei: We chose to develop on the Dotsama ecosystem because the substrate framework made it very easy for the initial integration. Substrate is a blockchain framework that allows you to build a very sophisticated blockchain network, somewhat similar to the Cosmos SDK.
We also needed to prioritize testing the archival storage within the Dotsama ecosystem using Substrate. However, this does not indicate us being limited to the Polkadot-Kusama ecosystem because we were already using Substrate. Our next step is integrating with the Ethereum ecosystem and enabling data availability and a bridging solution between the two.
vVv: Seeing what is being built by SSN is innovative and new. What challenges and hurdles does Subspace see regarding tech-security risks in the space when exposed to more commercial demand, user growth, and certain security risks?
Wei: One axiom is that no software is 100% secure. There will always be new bugs and exploits not by the logical implementation of the software itself but at lower levels like libraries, hardware, integration,.etc. This is going to be a constant monitoring and auditing process. The team is very opinionated regarding the quality of code testing. But we also wanted to ensure that we have a redundant auditing process. We’re not just going to rely on one auditing firm to confirm that the software or the consensus product is secure. We’ll probably have 3-5 independent auditors for each phase to ensure secure deployment. The process will apply to cross-chain and bridges.
vVv: Has Subspace been subjected to any external auditing?
Wei: Yes, we are already engaging with multiple parties, not just one. We’re even doing a security audit with our Farmer Node on the desktop app because we believe the desktop app is going to be the most widely adopted once we’re ready to launch. We have made it very simple for non-savvy crypto users to run the app on their computers and participate.
vVv: Subspace is described as a fourth-gen blockchain with no compromises. How does your L-1 solve the blockchain trilemma?
Wei: I think it’s more a question of how much a network can scale once they solve the initial trilemma because there will likely be future “trilemmas.” We solve the initial trilemma between security, scalability, and speed via two main adoptions. One is our development of a novel PoAS consensus model, a variant of the PoC model. That, combined with a decoupling execution model, allows us to have massive participation on the consensus side. Imagine 100,000s of farmers that will handle the security part, and because of the decoupling, each of them isn’t required to execute the same logic, which contributes to horizontal scalability. Then with the execution nodes, we can have much more of a specialized hyper-performance server setup where the farmers take care of the consensus and the ordering of the transactions in packaging the block content. Then we’ll ship it to the executor nodes, which will simply compute the state change. The beauty is that the executor node will only add deterministic transactions, which means anyone can check if a segment or portion of the transaction is valid or not. Just looking at the input and output is entirely deterministic. This means that we know an executor node might be highly performant in computing 100,000s of transactions. Still, any single node with a farmer or an external node can inspect a subset of those transactions to detect any error fault. Combining those two allows you to have a highly decentralized, highly performant, and scalable network. In the coming Gemini III, the distributed storage network, we will enable a farmer to forgo downloading the entire history but only manage a segment of that, making farming even simpler.
”We solve the initial trilemma between security, scalability, and speed via two main adoptions… our development of a novel PoAS consensus model …..combined with a decoupling execution model [which] allows us to have massive participation on the consensus side.
vVv: Currently, the Ethereum archive nodes require several terabytes of storage. Assume the Polkadot parachain ecosystem grows enormously so that the whole blockchain database is 100 terabytes. Will the protocol function properly even under this increase in demand? Will the system function sufficiently so that specific data calls will be answered within the same or the next block? What can Subspace nodes maintain modestly?
Wei: That’s a very technical one. I don’t have the full context on the efficiency of retrieval for arbitrary data. You have to understand that Subspace data is permanent storage, and if you have permanent storage, it’s never going to be as performant as more ephemeral ones. We believe there’ll be additional services being built on top of it. For example, cacheing, distributed networks, storage, and all those things that Web2 already figured out how to make highly scalable. In terms of the capacity to store multiple chains. For example, on Gemini I, we had 4-5 petabytes of storage pledged, meaning it can store hundreds if not 1000s of the Ethereum full archival nodes as history.
This means if there are 1000 chains as big as Ethereum, we can still store all of them on the Subspace Network and have excess space for other things. For PoC chains, supply is not an issue, and it’s more the demand, quality, and utility of the data being stored. On that front, Subspace is not just an arbitrarily decentralized storage platform, which happens due to the consensus model we have built. You can see the storage layer as more of a future than just an end.
Through the additional data service, data availability will be the trustless bridging to make the blockchain space complete. Initially, we were going to focus more on the quality of data rather than the quantity, but even with that said, we will have more than enough capacity for the entire blockchain ecosystem.
vVv: How do you assure permanent accessibility?
Wei: The beauty of the PoAS and the difference between Subspace and decentralized storage solutions like Filecoin or Areweave is that every piece of data stored on Subspace will be part of its history, and there’s no way to lose that unless the entire network goes down. It’s similar to storing data on the Ethereum network, which will be very expensive because it will be part of the system forever and has high-level guarantees and permanency.
Each node does not need to store the complete history, and that’s the beauty of enabling horizontal scalability of the network. Every piece of data stored on Subspace will go through ratio encoding, which means data is divided into multiple chunks and encoding them to multiply. Imagine if the information is divided into three chunks, and you encode with the output of nine chunks. This means any combination of three chunks can recover the original data.
The replication factor will be the security model beyond the fact that your data is saved in history. If everything is normal, you will never lose it. The replication factor determines data progression and the minimum data retention value. The good thing is that the replication factor is not manually determined and is semi-automatic via a market maker model. When additional nodes and space are pledged, that space will be empty, which will now store a segment of history that increases the replication factor. With the increase in demand, the replication factor decreases, and as a result of redundancy, the space is replaced with new data. All those fundamentals are meant to ensure extremely safe storage.
vVv: Arweave is the golden child of Web3 and blockchain and the preferred long-term decentralized storage solution. How do you position Subspace versus Arweave? What are some of Arweave’s shortcomings that Subspace solves?
Wei: I want to start by saying that we’re not competing with Arweave directly. Arweave provides the needed service of permanent decentralized storage for arbitrary objects or websites. However, Subspace is more than just storage; it is about what we store and the quality of data, rather than just arbitrary storage. There are some differences at the direct storage layer. The main difference is the Subspace consensus protocol built from the ground up to ensure that the storage layer is fully coupled with the consensus. Storage space, pricing, and the replication factor are dynamic. Any stored object is part of history and is very beneficial.
Arweave has a different notion between storage and proof of work (PoW). They have a hybrid system where the storage interfaces with a PoW mechanism. Another point is that the network’s total available space at any time is unknown. As a result, they employ a fixed price across the entire network per megabyte, which requires manual adjustments. In addition, the node or storage provider is not as dynamic as Subspace. There are quite a few fundamental differences on that front. Our goal is to make storage much more straightforward.