This episode reveals how DoubleZero (0ero) is tackling the critical bandwidth bottleneck in blockchains, offering a purpose-built internet to unlock new levels of performance and scalability for decentralized systems.
Episode Introduction
This episode dives deep into DoubleZero (0ero), a project building a dedicated, high-performance network layer to solve the often-overlooked bandwidth limitations plaguing blockchains, and what this means for the future of decentralized infrastructure and investment.
Introducing DoubleZero (0ero) and its Founders
- Austin, a co-founder of DoubleZero (0ero) attached to the foundation side, brings his experience from four years at Solana Labs and the Solana Foundation.
- Matteo Ward, co-founder of Malbec Labs, serves as a core contributor to the DoubleZero (0ero) protocol and network, focusing on protocol and network engineering.
- Matteo describes DoubleZero (0ero) as "an alternate internet, a purpose-built internet for distributed systems." This means it's specifically designed for the needs of time-sensitive, globally distributed systems requiring high performance, unlike the general-purpose public internet.
The Real Bottleneck: Why Bandwidth, Not Compute, Limits Blockchains
- Matteo explains that historically, tech platforms like YouTube and cloud providers (AWS, Google Cloud) eventually built private, purpose-built networks to achieve high performance and scale, a step he argues blockchains now need to take.
- He states, "It's no magic. It's just what did every other tech platform do when they needed to become high performance and scale? They built a purpose-built custom dedicated network underneath."
- Austin challenges the common focus on compute as the primary blockchain bottleneck. He points out that testnet performance often far exceeds mainnet performance due to controlled environments and smaller scale, not just different load types.
- Modern validator clients, like Solana's Fire Dancer (capable of a million transactions per second in controlled tests on the DoubleZero network), demonstrate that software is no longer the main constraint. The public internet's limitations hinder this potential in real-world, global deployments.
- Strategic Implication: Investors should recognize that improvements in raw compute power for validators might yield diminishing returns if network infrastructure doesn't keep pace. The focus is shifting to the data transmission layer.
How the Public Internet Fails Blockchains
- Austin highlights the inefficiency of current internet routing. Traffic between, for example, New York and Singapore, often takes convoluted paths due to business models prioritizing least-cost routing and quick handoffs, not performance.
- There's no mechanism to pay for packet prioritization on the public internet.
- Matteo elaborates on the technical reasons: the internet uses many Autonomous System Numbers (ASNs)—essentially different networks—that pass traffic using Border Gateway Protocol (BGP). BGP routing decisions are influenced by various factors, including cost and geopolitics, leading to inefficient paths similar to airline hub systems.
- ASN (Autonomous System Number): A unique number that identifies a distinct network on the internet, managed by a single administrative entity.
- BGP (Border Gateway Protocol): The routing protocol used to exchange routing and reachability information among ASNs on the internet.
- DoubleZero (0ero) addresses this by operating as a single global ASN. This allows direct, optimized routing between any two nodes on its network, creating a fully peered mesh.
- Matteo: "What the magic of what DoubleZero (0ero) does is it routes the entire world on a single ASN number, which means that every single packet goes straight to the location that it's going to."
- This decentralized, single ASN approach, built with multiple network providers, is what distinguishes DoubleZero (0ero) from centralized private networks.
The Role of Tokens and Network Contributors
- Austin emphasizes that tokens are crucial for incentivizing all participants in the network chain, ensuring that each network provider facilitating an efficient route gets compensated. This aligns incentives towards performance, unlike the current internet.
- Links providing the biggest advantage over the public internet and reaching the most stake/validators/RPCs are rewarded most.
- Matteo outlines the stakeholders:
- Early Adopters: Entities understanding blockchain, utility tokens, possessing underutilized high-performance network capacity, and willing to take risks (e.g., Jump Crypto, Rockaway X, Distributed Global).
- Phase Two (Wait and See): Carriers and enterprises who observe initial success and can build a business case for participation.
- Phase Three (Entrepreneurs/Capitalists): Individuals or groups who can assemble the necessary data center space, hardware, and connectivity to contribute to the network and earn tokens.
- Austin notes the professionalization of blockchain infrastructure. While individuals and small firms can contribute, it's a serious commitment, not a home-user hobby, reflecting the maturity and profitability of the sector.
- Actionable Insight: The DePIN model of DoubleZero (0ero) relies on professional network operators. Investment opportunities may lie in entities that can efficiently provide and manage this specialized infrastructure.
DoubleZero (0ero) Tokenomics and Revenue Model
- Austin directly addresses a common issue with DePIN (Decentralized Physical Infrastructure Networks) projects: the lack of a demand side. DePINs often incentivize supply without corresponding usage, leading to token value decline.
- DePIN: Networks that use token incentives to bootstrap and operate physical infrastructure like wireless networks, storage, or in this case, specialized internet backbones.
- DoubleZero (0ero) aims to avoid this by:
- Not rewarding mere passive fiber contribution; rewards are tied to the usefulness of the fiber in connecting active network participants (validators, RPCs).
- Charging validators and RPCs (Remote Procedure Call nodes) for using the DoubleZero (0ero) network, creating revenue inflows. RPCs are servers that allow dApps to interact with the blockchain.
- A portion of collected fees (paid in native tokens like SOL or DoubleZero (0ero) tokens, then converted to DoubleZero (0ero)) is burned, and a portion is distributed to network contributors. This is revenue-derived, not purely inflationary.
- Austin: "The problem with DePIN is not that they pay out in the local currency. The problem with DePIN is there's usually no demand side."
- The focus is on utility: if DoubleZero (0ero) makes networks like Solana significantly more performant and profitable, the demand and token value will follow.
Fee Structure and Value Proposition
- Austin discusses the proposed 5% take rate from validator rewards, acknowledging it needs to be earned by demonstrating significant value (e.g., making validators "twice as profitable").
- He strongly advocates for protocols to "turn the fee switch on" early to validate product-market fit, arguing that free services don't prove genuine demand.
- Matteo frames DoubleZero (0ero) as a blockchain-native network offering the most efficient way to "mine bandwidth." It allows participants to extract bandwidth at marginal cost without large upfront capex, sharing infrastructure and governance.
- Strategic Consideration: The success of DoubleZero (0ero)'s token hinges on its ability to demonstrably improve the economics for its users (validators, RPCs). Researchers should monitor adoption rates and reported performance gains.
Performance Improvements and Technical Mechanisms
- Matteo explains the tech stack:
- Utilizes advanced low-latency switches and high-quality, low-jitter fiber routes (preferably wavelength-based).
- Implements a filtration stack with FPGAs (Field-Programmable Gate Arrays), specifically AMD V80 cards. FPGAs are integrated circuits that can be configured by a designer after manufacturing, offering high performance for specific tasks.
- These FPGAs act as a "bump in the wire," efficiently discarding improperly signed or duplicate packets before they reach validators, freeing up validator CPU.
- Expected improvements:
- Potentially a 70% reduction in "garbage traffic" that validators currently process.
- 30% to 200% aggregate latency reductions, with the wide range due to the internet's variability.
- Matteo mentions a recent testnet upgrade to IBRL mode (IP-Based Routing Layer), which will provide more stats on these improvements.
- IBRL (IP-Based Routing Layer): A mode of operation for DoubleZero (0ero) focusing on optimized IP routing, potentially for networks that don't require the intensive filtration services.
Understanding "Garbage Traffic" Filtration
- Austin clarifies that "garbage traffic" includes:
- Transactions failing signature verification: Packets where the cryptographic signature doesn't match the purported sender, often sent maliciously to clog networks (economic denial of service). FPGAs can perform signature verification at 100 gigabits per second.
- Duplicate packets: Multiple copies of the same packet sent to overwhelm validators. FPGAs can perform packet deduplication at high speeds.
- This filtration provides validators with a cleaner stream of potentially valid transactions, though not all will be high-value or succeed in execution (e.g., a trade at an outdated price).
- Future FPGA capabilities could include flagging potentially high-value transactions based on criteria like CU-to-fee ratio, aiding validator profitability.
- Austin clarifies that FPGAs are not suited for stateful operations like on-the-fly auction mechanisms, which would add latency. Such services (like Jito's) would be built on top of DoubleZero (0ero) using traditional compute.
- The 100 gigabits per second processing capacity of the FPGAs is 10-100 times what most validators currently use (1-10 Gbit connections) and exceeds the largest DoS attack recorded on Solana (80 Gbps).
Expansion to Other Chains and Use Cases
- Austin states Solana currently has the highest product-market fit due to its high-capacity validator client, large validator/RPC set (approx. 2000 nodes), and highly competitive activity (e.g., for blockspace).
- Networks without intense competition for blockspace (e.g., DA layers like Celestia, storage protocols like Filecoin) may not immediately need the full filtration suite but can benefit from the IBRL mode for optimized routing.
- Austin: "If we can nail the Solana use case, expanding out to other networks... the rest is downhill from there."
- Competitive use cases imply significant monetary value in being first to get a transaction into a block, driving demand for performance-enhancing infrastructure like DoubleZero (0ero).
The Power of Multicast
- Austin explains that networks often reduce validator counts to achieve speed because propagating data to many nodes is exponentially harder. DoubleZero (0ero) aims to solve this with multicast.
- Multicast: A communication method where information is addressed to a group of destination computers simultaneously, sending a single stream of data that is then replicated by network routers only when paths diverge to reach multiple recipients.
- Matteo elaborates: Multicast is how stock markets efficiently propagate market data. However, it requires a single ASN, which the public internet doesn't offer for this purpose, but DoubleZero (0ero) does.
- This allows efficient traffic propagation even with large, globally distributed validator sets, countering the trend of shrinking validator sets for performance.
- Matteo: "We're right now in testing mode on multicast with Jito and it's going well. It works."
- Research Note: Multicast capability could be a game-changer for maintaining decentralization while scaling performance, a key area of research in blockchain architecture.
DoubleZero (0ero)'s Identity: Beyond DePIN to "N1"
- Austin notes that DoubleZero (0ero) differs from typical DePINs:
- Most DePINs offer lower quality but higher reach (e.g., Helium vs. cell towers).
- DoubleZero (0ero) offers much higher quality than the public internet but doesn't have its ubiquitous reach (it's not a residential service).
- Matteo introduces the "N1" (Network Layer 1) concept: DoubleZero (0ero) is a base layer of infrastructure, operating at layers 1-3 of the OSI model, while blockchains are typically at layers 4-7.
- Strategic Implication: Viewing DoubleZero (0ero) as an "N1" positions it as foundational infrastructure upon which other "N2" services and applications (including AI coordination, precision time, etc.) can be built, creating a new ecosystem.
Future "N2" Applications
- Matteo envisions an ecosystem of "N2s" built on DoubleZero (0ero), including:
- File storage solutions.
- Coordination of AI training and inference models.
- Precision time services.
- Geo-awareness capabilities.
- These N2s are crucial for advanced blockchain features like multiple concurrent leaders, which require precise timing and location data.
- Crypto AI Relevance: The potential for N2s focused on AI model coordination and inference directly connects DoubleZero (0ero)'s infrastructure to the Crypto AI space, enabling more performant and decentralized AI applications.
Retrospective Learnings
- Austin admits underestimating the activity levels on chains other than Solana. The initial thesis was that many high-performance networks would immediately need DoubleZero (0ero)'s filtration services. In reality, fewer chains currently have the throughput and competitive intensity to fully leverage them, leading to prioritizing the IBRL mode for broader applicability.
- Matteo would have aimed for a 100 Gbps backbone from the start, rather than beginning with 10 Gbps, as market demand and contributor capacity for 100 Gbps emerged faster than anticipated.
- Matteo: "If I did it again I would go directly to 100 gig. I'd be even more bullish on it."
Future Outlook (6-18 Months)
- Austin hopes for DoubleZero (0ero) to be on mainnet within 6 months. Over 18 months, the goal is to provide capacity for networks to significantly scale their available resources.
- He anticipates a "high-performance blockchain renaissance," with increased market competition on performance and greater specialization, enabled by a non-commoditized, high-performance network stack like DoubleZero (0ero).
- This could lead to new blockchain architectures, possibly fully hardware-accelerated, leveraging DoubleZero (0ero)'s capabilities.
- Matteo looks forward to widespread adoption and traffic. By Q3, he expects significant validator and stake onboarding, with a mainnet beta launch. By Q4, he anticipates demonstrating major aggregate performance bumps as validators natively use the network.
- Matteo visualizes the network evolving from "hairpinning" traffic through the public internet to native transit and efficient multicast distribution within the DoubleZero (0ero) core as adoption grows.
Conclusion
DoubleZero (0ero) is positioning itself as a fundamental enabler for the next generation of high-performance blockchains by directly addressing network bandwidth and latency. Crypto AI investors and researchers should monitor its adoption by major L1s/L2s and the development of "N2" applications, particularly those facilitating decentralized AI.
