DoubleZero: A Faster Internet
Introduction
The journey begins with a single flicker. A spark of light is born amidst the vastness of a data center’s humming machinery. A packet—a single transaction, is encoded into tiny bursts. These pulses begin their journey through a fiber of optical filaments, each finer than a strand of human hair, stretching across the vast Atlantic Ocean. Deep beneath the surface, this cable of fibers follows the contours of the seabed, weaving through miles of crushing pressure and total darkness. Yet, the light pulses onward with unwavering precision, crossing continents in silence. This journey is happening millions of times a second as you read this. It is the hidden backbone of global connectivity—the unseen infrastructure that bridges humanity’s digital dreams, one unstoppable pulse at a time.
DoubleZero aims to redefine this unseen connectivity backbone, laying the foundations for a faster global fiber optic network—one that will meet the demands of a new generation of high-throughput, low-latency distributed systems and blockchain ecosystems, including Solana. This addresses a fundamental bottleneck for both the Internet and high-performance blockchain systems—global, base-layer connectivity.
DoubleZero’s ‘N1 protocol’ introduces neutral, base-layer infrastructure tailored for a high-performance distributed future. The protocol envisions a permissionless network powered by independent contributors, leveraging dedicated fiber-optic and subsea cables to re-engineer the foundational OSI model layers (Layers 1, 2, and 3).
This article delves into the core insights from the newly released DoubleZero whitepaper, highlighting its market potential, high-level system architecture, and economic model. DoubleZero has the potential to deliver substantial performance gains for networks such as Solana. However, It is fundamentally blockchain-agnostic and designed to provide transformative benefits across various industries and use cases. This versatility will be explored in detail throughout this article.
Market Opportunities
‘Dark fiber’ refers to unused fiber-optic cable that has been installed but is not being used. Its opposite is ‘lit fiber’—literally fiber with light traveling through it—which refers to fiber-optic cable in use. Organizations can lease and manage dark fiber to create a private network, as well as lease space on existing lit fiber.
Supply Side
Substantial spare capacity dark fiber has been installed but is not currently operational.
- 65% of US-based fiber remains dark (Federal Communications Commission, 2007)
- Fiber capacity rose from 15% to 46% of US locations from 2015 to 2023 (Fiber Deployment Annual Report, 2023)
- 70% of European households have fiber connections (FTTH Council Europe, 2024)
Fiber link providers with underutilized or disconnected links face significant challenges in monetizing them individually. Selling services for each route piecemeal involves high overhead for finding buyers and negotiating contracts, while the value of each link remains limited to its specific route. Many enterprises tend to overprovision capacity when buying or leasing fiber, building it for worst-case rather than routine outcomes.
Demand Side
Data flow bottlenecks constrain the performance of decentralized networks. In high-performance blockchains, bandwidth limitations and unpredictable communication latency between validators greatly outweigh computational constraints. Public internet routes exacerbate these issues with non-deterministic routing. A blockchain's dual objectives of being public and performant often conflict.
Validators must contend with a relentless influx of inbound transactions riddled with duplicates and overwhelmed by spam. Even the most performant blockchains struggle to deliver the Web2-like user experiences users expect and the trading experience offered by professional centralized exchanges.
The industry has introduced several solutions to address these challenges, each with its limitations:
Mempools: Mempools act as centralized holding areas for transactions, reducing the ingestion burden on individual validators. However, they introduce additional latency and broadcast transaction intentions well before their inclusion, potentially exposing them to exploitation.
Quality-of-Service (QoS): Stake-weighted QoS systems prioritize transactions from trusted entities with stake. Unfortunately, these systems add extra latency due to the proxy step of running transactions through a staked node. Participants without sufficient access to stake are excluded.
DoubleZero’s Solution
DoubleZero envisions creating a decentralized physical infrastructure network (DePIN) of underutilized private dark fiber links and contributions from large network operators like financial firms and telcos. Essentially, it is a meshed transport layer that filters and serves traffic across low-latency and high-bandwidth routes, allowing distributed systems to send and receive information efficiently. Individually contributed fiber links are combined into a synchronized network that increases bandwidth, lowers latency, and removes jitter from communication. This dedicated bandwidth provides predictable, fast connections between network nodes.
To illustrate, imagine the difference between relying on Uber to travel from point A to point B and having a personal driver. A private driver is always available, providing immediate and predictable fast service. Relying on Uber could be more flexible, but it will introduce variances in arrival time and potential delays depending on driver availability.
Yet DoubleZero is not just a transport layer—the network also provides filtration and verification services conducted by network devices running open-source code to verify signatures and deduplicate inbound data. Inbound transactions can be edge-filtered (i.e., remove spam and dedup) by specialized hardware. This allows blockchains to benefit from shared system-wide filtration resources rather than needing each validator to provision sufficient resources. Validators receive a substantially smaller transaction set and must only re-verify signatures on the transactions included in the final block. Outbound messages can be explicitly routed, tracked, and prioritized to improve efficiency.
Under this system, we no longer need to provision individual validators with enough resources to meet global demand; instead, this demand can be met through infrastructure sharing, requiring fewer resources in aggregate. Custom infrastructure will be used at the fiber endpoints, so traffic routing, global latency, and other lower-level performance will be superior to the public internet.
DoubleZero’s core focus is on fiber, whether terrestrial or subsea. Modern fiber links can support hundreds of terabits of data per second. However, network contributors can also provide other communications services, such as microwave or satellite technology, and attach them to the network as an ‘N2’ application.
Architecture
You can conceptualize the architecture of the DoubleZero network as two concentric rings.
Outer ingress/egress ring: Interfaces with the public internet. DoubleZero’s network devices use performant hardware to mitigate distributed denial-of-service attacks, verify signatures, and filter duplicate transactions.
Inner data flow ring: Servers on the inner ring build consensus with this filtered traffic over optimally routed dedicated bandwidth lines.
The physical layer of the DoubleZero network consists of network devices at key ingress/egress points and provisioned bandwidth across the network.
DoubleZero Exchange Points (DZXs)
DZXs enable connectivity between multiple data centers within a metro area, similar to how public internet exchanges connect different networks. Network contributors provide these underlay links, which are the protocol's core transit layer. The exchange points allow scalable interconnectivity, regardless of how many contributors participate in DoubleZero.
Fiber Links
Fiber links provide low-latency, high-bandwidth connections between different points on the globe. The links can be directly owned by the network contributor or leased by them from the major network service providers. Network contributors provision these connections and commit to a service level agreement for each connection’s quality.
FPGA Network Devices
Field-programmable gate arrays (FPGAs) are hardware devices that can be programmed after manufacturing to perform specific computational tasks. Known for their flexibility and high performance, FPGAs are commonly used in applications requiring rapid data processing, such as networking, signal processing, and blockchain systems.
DoubleZero uses commercially available FPGAs to filter inbound traffic at key ingress points at the network edge. Compared to validators, FPGA appliances can handle orders of magnitude more traffic as they perform the specific tasks of removing spam, de-duplicating a transaction set, and verifying signatures. One sample deployment alone can handle deduplication, filtering, and signature verification for multiple Gbps of inbound data: up to a million transactions per second with zero latency.
Scenario: DDoS Attacks
DoubleZero’s infrastructure sharing creates a more robust and resilient network architecture, making a denial-of-service (DoS) attack much more challenging. Under DoubleZero, such an attack would require many terabits per second of traffic, simultaneously hitting many geographically distributed sites. It would no longer be an attack against an individual validator or even a single blockchain; it would be an attack against hundreds of data centers and ISPs worldwide. A successful attack would have to be several orders of magnitude more complex and resource-intensive.
Economic Model
“We need good mechanisms and rich incentives to ensure that network contributors want to add their underutilized channels and distributed systems want to use the resulting network. In short, we need a new economic model for communication.”
- Nihar Shah, Chief Economist, DoubleZero Foundation (source)
DoubleZero introduces a novel economic model for filtration and routing services across global networks, connecting underutilized fiber resources with decentralized systems. It transforms individual contributions into an integrated network, creating a system more valuable than the sum of its parts. Even a single, redundant fiber link—limited in standalone value—becomes a critical component when combined with others in a distributed framework.
DoubleZero aggregates these diverse contributions, blending the strengths of private networks (unified standards but limited reach) with those of public networks (broad reach but fragmented standards).
Fiber Suppliers
DoubleZero enables private businesses to monetize underutilized fiber links they have purchased or leased from telecom providers or network operators. By contributing these links to DoubleZero, businesses unlock new revenue opportunities.
Blockchain Demand
Blockchain networks can access the advantages of private networking—such as enhanced performance and reliability—without dependence on centralized systems or long-term contracts.
With DoubleZero, users provision and utilize fiber seamlessly through blockchain-based smart contracts, ensuring the integrity and transparency of every interaction.
Smart Contract-Defined Network
DoubleZero operates using verified data stored in on-chain smart contracts. A permissionless controller deployed on a public blockchain incentivizes new contributions and dynamically manages network configuration and routing to address demand spikes, outages, and other disruptions.
The distributed control plane leverages this data to reach consensus on routing and prioritization decisions, ensuring a transparent and permissionless experience for contributors and users.
Contributors connect to the network by establishing a service-level agreement encoded in a smart contract. This agreement specifies key link parameters such as endpoint locations, bandwidth, latency, and compliant MTU size. For instance, a contributor providing a link between Los Angeles and Singapore might commit to delivering 10 Gbps bandwidth with 85ms latency and a 1600 MTU. These details and the duration of the commitment are recorded in the smart contract, defining the link's operational characteristics.
DoubleZero monitors and tests link performance to ensure only compliant links qualify for rewards. Non-compliant links risk disqualification from rewards and may be removed from the network. The protocol relies on internal alignment and reciprocity among contributors, who must adhere to service-level agreements and remain incentivized to hold each other accountable.
Validators and other users can opt into predefined services and routes—such as traffic for a specific Layer 1 blockchain, state propagation, or validator sets—or author custom contracts to define specific routing requests.
The distributed controller uses smart contract data as input to determine the network's optimal configuration. A key aspect of this process is calculating accurate pricing, ensuring users pay proportionally to the demands their requests impose on the system. Providers, in turn, are rewarded based on the value their contributions bring to the network.
DoubleZero envisions a tiered pricing model, allowing users to choose between priority and base access for specific routes, optimizing network efficiency. By offering a high level of customization and control over its services, DoubleZero ensures that value is passed back to providers—a significant advantage over the public internet, where all traffic is treated uniformly regardless of priority.
Network Effects
The network effects of DoubleZero resemble those of the airline industry. Due to limited demand, individual airline routes between two cities are rarely offered in isolation; instead, they are integrated into a carrier’s larger network, connecting through hubs and interacting with other routes. These networks become more valuable when carriers join alliances, extending their reach globally and amplifying the value of each route.
By contributing their fiber links to DoubleZero, providers unlock two major benefits:
- Ease of Integration: Adding links to the platform is simple and standardized. Providers only need to install and configure a few switches, keeping overhead minimal.
- Increased Link Value: A contributed link gains value through network effects by connecting with neighboring links, leveraging Metcalfe’s law. A previously isolated link can transform into a highly connected asset within DoubleZero, becoming far more valuable than it would be as a standalone route.
DoubleZero can even effectively utilize redundant, short, or low-performance links. While the network is designed to be high-performance and expansive overall, these contributions enhance redundancy, increase capacity, and improve overall interconnectedness.
Blockchain Use Cases
DoubleZero’s blockchain-specific applications include powering RPC nodes, optimizing MEV systems, and enhancing the performance of both Layer 1 and Layer 2 blockchains.
L1 Blockchains
Filtering: DoubleZero filters inbound traffic, helping to mitigate distributed denial-of-service attacks, verify signatures, and remove duplicate transactions. Separating these steps from block production, execution, and consensus frees up substantial resources to focus on block production.
Routing: Done through dedicated links with minimal latency and jitter. This ensures that blocks are built, shared, and ratified quickly. DoubleZero is expected to support multicast, facilitating simultaneous one-to-many or many-to-many data transmission. Multicast allows for the complex propagation of state transitions, whole blocks, or, in Solana’s case, shred streaming across the Turbine tree.
Traceability: Packets traversing over the network are expected to be traceable so that users can verify their inclusion or exclusion from intra-validator communication. This is a beneficial by-product of transparency in traffic routing for systems sensitive to censorship concerns.
RPC Nodes
RPC nodes face three challenges that DoubleZero could mitigate.
Traffic Surges: RPCs are especially vulnerable to distributed denial-of-service attacks because they receive transaction flow via the public internet. Canonical examples include airdrops and hot NFT mints.
Transaction Forwarding: Transaction deliverability matters immensely for an RPC, especially for high-value transactions that are both competitive and time-sensitive. Resilient and low-latency connections from the RPC node to the block leader or mempool assist RPC nodes in these responsibilities.
Information Retrieval: RPCs are primary providers of observability about the current state, including wallet balances and trading liquidity. They must ingest, validate, and index changes as close to real-time as possible. DoubleZero’s resilient and low-latency connections from validators to RPC nodes can ensure that RPCs have up-to-date views of the blockchain.
MEV
MEV systems are latency-sensitive; even minor improvements in latency and deliverability can deliver immense additional value. These systems benefit greatly from a more real-time view of a network's state. They can construct profitable transaction orderings and craft more efficient blocks with fresher data. MEV systems that relay proposals to validators quickly can use the extra time to find better orderings. MEV systems benefit from faster data retrieval (i.e., read operations) and quicker delivery of blocks or bundles to the current leader (i.e., write operations).
Layer 2 Blockchains
Layer 2 chains can utilize DoubleZero by filtering inbound transactions to the sequencer, posting to or retrieving information from Layer 1 with low latency, and quickly verifying updated state across the validator set. DoubleZero can also ensure that the data availability layers keep pace with the Layer 2 chain.
Most Layer 2 chains aspire to support multiple sequencers posting transactions in rotation or tandem to reduce centralization, censorship, and liveness risks. This architecture shift
requires substantial coordination between sequencers, which DoubleZero can speed up.
Further Blockchain Contexts
DoubleZero offers advantages in other blockchain contexts, including state synchronization,
indexers, multiple leaders, and network extensions.
New Validator Onboarding: To catch up to the current state, new validator nodes must ingest transactions substantially faster than the blockchain's current speed. DoubleZero's high-bandwidth connections help overcome these constraints and make it easier for new validators to join a network. The same principle applies to validators who have fallen behind the tip of the chain and need to catch up.
Indexers: Like RPCs, indexers must receive updated state fast so their end users can make quick, informed decisions. State propagation through DoubleZero reduces delays in indexing fresh data and keeps distributed indexer instances current.
Multiple Leaders: Many Layer 1 blockchains aspire to have multiple concurrent leaders simultaneously produce blocks in different parts of the world. Without tight coordination between leaders, this can quickly lead to redundant or outdated transactions and wasted compute. Like multiple sequencers on Layer 2 chains, multiple leaders on Layer 1 chains need rapid global state synchronization, and DoubleZero can facilitate this better than the public internet.
Network Extensions: Nearly every network has the concept of a compute operation that can be run on only a subset of a network’s validators as it requires specialized hardware or software. DoubleZero's increased bandwidth and reduced latency links can benefit such projects, including ZK provers, GPU workloads, and asynchronous execution platforms.
Non-Blockchain Use Cases
DoubleZero is not limited to blockchain applications. It can benefit any setting where distributed systems communicate quickly, precisely, and voluminously. Examples include content delivery networks, gaming applications, large language models, and enterprise use cases.
Content Delivery Networks (CDNs): A new CDN built atop the DoubleZero network could focus on the core competencies of provisioning content based on demand forecasting, leaving the filtering and routing to the protocol.
One major advantage of the DoubleZero protocol is its flexibility. Users do not need to contract for expansive and long-term leases but can customize their usage for specific routes and time intervals. Thus, CDNs that lack coverage in particular regions or need more bandwidth across central corridors can use DoubleZero accordingly.
Online Gaming: Latency is critical for online games, with even mild lags ruining the user experience. Game developers use two models: peer-to-peer and central servers. DoubleZero, with its dedicated links and flexible usage model, can augment either by stabilizing connections and making games more enjoyable.
Large Language Models (LLMs): Co-locating large GPUs in a single facility is challenging. Large tech firms have designed systems to coordinate asynchronous and distributed resources such that they can train unified models. DoubleZero’s high-bandwidth connections enable models to be trained more seamlessly across various facilities and GPU availabilities.
Enterprise Use Cases: DoubleZero can offer an alternative to standard cloud provider offerings, allowing users to purchase and utilize bandwidth permissionlessly without locked-in leases or contracts. Further advantages include cost reduction, alleviating data center ingress/egress congestion, and avoiding ecosystem lock-in.
Conclusion
DoubleZero is one of the most ambitious undertakings in the blockchain industry today. The vision is to build a new private internet tailored for modern distributed systems—including Solana—that seeks to redefine the infrastructure powering decentralized technology. Led by former Solana Foundation Head of Strategy Austin Federa, CTO Andrew McConnell, and data networking expert Mateo Ward, DoubleZero launches with the backing of two core contributor teams, Firedancer and Malbec Labs. Testnet is live in seven cities globally, with mainnet to add an additional ten metro areas. Four Solana validators (two Agave and two Frankendancer) are connected to the DoubleZero network today in a test capacity. They are hiring.
Thanks to 0xIchigo, Brady Werkheiser, Andrew McConnell, and Austin Federa for reviewing earlier versions of this work.
