### GridPool 201: Advanced Architectural Deep Dive #### 1\. Protocol Philosophy: From Centralized Accounting to Shared Lottery Mining GridPool represents a fundamental shift in mining architecture, moving away from the systemic centralization risk arising from reward concentration in custodial pool wallets. In the current landscape, Full Pay Per Share (FPPS) pools dominate by absorbing variance, yet they require massive hashrate to remain solvent and often dictate block construction, creating a "laughable" but dangerous centralization of transaction selection. Several technical solutions have been put forward to decentralize template construction, but by attempting to replicate traditional share accounting each introduces vulnerabilities. GridPool is engineered not as a pool, but as a peer-to-peer protocol enabling "Shared Lottery Mining." It occupies a unique technical niche, making certain engineering tradeoffs that give it distinct advantages over other attempts to decentralize the mining process. * **Versus sharechain based decentralized pools:** GridPool eschews the maintenance of a secondary blockchain (sharechain), avoiding the associated computational overhead, latency-induced centralization pressure, and vulnerability to 51% attacks on the sharechain. * **Versus Ocean and Stratum V2:** While Ocean and Stratum V2 have spearheaded decentralized template construction, both maintain a centralized entity for share approval and payout determinations. GridPool accepts higher variance in exchange for client-side consensus based reward distribution without a central reward-collection wallet or administrative intermediary, reducing censorship risk. * **Versus pure solo mining:** Achieves up to 300x reduced variance compared to pure solo mining through reward sharing. Small miners benefit more than large miners, which further promotes decentralization of hashpower (a large miner with 50% of the pool’s total power only realizes a 2x variance reduction). **Core Protocol Advantages:** * **Simplicity and scalability:** Compared to other technical solutions, this protocol is simple, new client and node implementations are straightforward to add. * **Optimized Overhead:** The removal of internal Directed Acyclic Graphs (DAGs) or sharechains minimizes both bandwidth requirements and verification latency. The bandwidth requirements can easily scale to accommodate the entire bitcoin network if needed. Nodes can optionally transmit new shares via a single UDP packet \<1200 bytes, an innovation borrowed from the FIBRE network. * **Regulatory Resilience:** By eliminating a central payout custodian, the protocol operates beyond the reach of standard "money transmitter" regulatory frameworks. * **51% Attack Resilience:** An attacker with overwhelming mining power cannot reorg or alter GridPool’s consensus mechanism. Even a 99% attack could only hope to claim 99% of the payout slots, which is the expected appropriate outcome * **Censorship resistance:** Shares include only a bare minimum of information to enable trustless verification by other nodes, and expressly do not include data on chosen transactions other than the coinbase transaction and the minimum required Merkle leaf. Other nodes cannot censor shares based on a miner’s chosen transaction set. * **Block Witholding Attack resistance:** In exchange for slightly increased variance, GridPool allows block finders to claim an immediate portion of the block reward, plus all transaction fees, incentivizing immediate submission of blocks. Miners that attempt to withhold blocks will be hurt more than honest miners. * **Fast payout snapshots:** In protocol V2, every new Bitcoin block turns the highest-ranked unpaid work into the active payout snapshot. A GridPool block pays that active snapshot, then only the paid proof IDs are removed from the unpaid reserve. This gives new contributors a faster path into live templates without requiring a special high-fee genesis round. * **Layered pooling:** The GridPool protocol can be used by most other mining pools to split rewards among themselves, reduce variance, and drastically reduce the minimum size required to remain viable. This presents a significant opportunity for small to medium sized pools. **Notable Disadvantages:** Like any engineered system, the above advantages come with tradeoffs, listed here for full transparency. * **Higher variance:** While up to 300x lower than pure solo mining, GridPool users will see notably higher variance than traditional FPPS or PPLNS pools. Compared to standard pools, Grid Pool does not track individual shares with as much fidelity, * **No transaction fee split:** GridPool makes no attempt to split transaction fees between participating miners. If or when transaction fees become a significant proportion of the block reward, GridPool miners will experience higher variance in payouts without protocol changes to enable zero knowledge proofs of proper fee splits. * **Cannot integrate some other pools:** While small to medium sized pools can use this protocol to their advantage, they cannot do so if they are designed to be fully non-custodial and to pay rewards directly from the coinbase reward. This is because GridPool uses the coinbase transaction in it’s consensus layer. #### 2\. Eliminating the Sharechain-Majority Attack Surface The architectural resilience of GridPool is defined by its departure from the sharechain model. The sharechain is a separate blockchain whose security is bounded by the hashrate participating in that specific pool. This creates a "pool-chain majority" attack surface where a powerful miner can reorganize the sharechain or rewrite history independently of the Bitcoin mainnet. GridPool eliminates this attack vector by removing the internal ledger. There is no historical "longest-chain" of shares to manipulate or privately mine. **Distinction of Adversarial Models:** * **GridPool Dominance:** An attacker with a majority of GridPool-specific hashrate can dominate future payout snapshots through proportional mining power—this is simply the intended function of proof-of-work (PoW) attribution. They cannot reorganize the payout distribution or censor other miners. * **Bitcoin-Level Security:** GridPool is not a separate chain; it is a collaborative coordination layer. An attacker cannot "reorg" GridPool history because there is no chain to reorg. GridPool remains Sybil-resistant because "identities do not vote"—shares are ranked strictly by verified cumulative difficulty. #### 3\. Censorship Resistance via Transaction Blinding GridPool employs a share-relay layer that blinds peers to the specific transaction set selected by a miner, thereby resisting transaction-level censorship. Relaying nodes verify PoW and payout commitments via a minimal data subset rather than the full block template.The technical implementation restricts data sharing to the block header, the coinbase transaction, and specific Merkle paths: GridPool peer share messages include the block header, coinbase transaction, and the Merkle path needed to prove the coinbase commits to the header Merkle root. They do not include the full transaction list or a Merkle proof for arbitrary transactions. See *MiningModels.cs* (line 8\) and *BootShareVerifier.cs* (line 156). **Technical Caveat:** While providing high resistance, this is not absolute blinding. The coinbase branch leaks a small number of Merkle subtree hashes, and depending on the ordering, the "first sibling" can reveal the transaction ID (txid) immediately following the coinbase. Technically, peers can attempt to censor visible metadata such as payout addresses, coinbase tags, or parent block hashes, however they risk splitting themselves off the main pool by doing so. See “5. Consensus Mechanisms: Dynamic Team Convergence” for more on this. #### 4\. Networking Stack: Multi-Layered Communication & Propagation To ensure high-speed share propagation and resilient control planes, the GridPool networking stack utilizes a three-tier hierarchy: 1. **V3 Relay (UDP):** The latency-critical tier used for high-speed share propagation across the P2P network. 2. **WebSockets:** The primary backup for reliable, stateful communication and peer-to-peer handshakes. 3. **HTTP:** The fallback layer for health checks, command-line interactions, and administrative probes.**Node Orchestration:** The default bootstrap seed is located at *gridpool.net* , facilitating initial peer discovery. Nodes expose a WebUI on port 5000 and a DATUM listener on port 3008\. Gossip mechanics are driven by the *NotificationSource* , which can be configured for "Zero-node" mode via *MempoolSpace* or sovereign mode via a local *bitcoind* ZMQ feed. #### 5\. Mitigating Block Withholding Attacks: The Slot-0 Advantage GridPool utilizes an incentive structure to solve the block withholding problem inherent in pooled mining. In any valid GridPool template, the miner who successfully finds the Bitcoin block is allocated "Slot 0” in the coinbase transaction, immediately winning one conceptual payout slot, any subsidy remainder, and all transaction fees. **The Protection Fee and Economic Game Theory:** * **Immediate Realization:** Unlike PPLNS or sharechain rewards, the Slot-0 reward is never promised or tallied long-term; it is realized only upon block submission. This provides an immediate, tangible incentive to broadcast. * **Fixed slot accounting:** The current beta uses 300 conceptual payout slots. Slot payouts are fixed at subsidy / 300. Slot 0 receives one slot's worth through the remaining coinbase value, plus the subsidy remainder and all transaction fees. With the default support fee enabled, one post-slot-0 output is the canonical Grid Labs support slot and up to 298 shared proof slots are paid. If the support fee is disabled, up to 299 shared proof slots are paid. * **Adversary Detection:** Nodes can detect non-contributing "attackers" by identifying nodes that consistently appear in high-difficulty payout snapshots but never submit a Slot-0 block over statistically significant periods. #### 6\. Client Ecosystem and Layered Integration (DATUM & Firmware) GridPool is designed as a base-layer protocol for the broader hashing ecosystem. It currently supports the **DATUM** gateway, with a roadmap including **Hydrapool** , **Stratum V2**, **CKPool**, **Public Pool**, **Esp-Miner, and Mujina.** **Simplified Client Integration:** Because the protocol itself is so simple to implement, further client integration is trivial. For example, experimental firmware has even been built for the Bitaxe, allowing ESP32 based desktop miners to construct their own block templates and submit shares directly to the GridPool network, with no Stratum layer at all. This firmware was built by an AI agent (Codex 5.5) in \~30 minutes from a single prompt. **Protocol Layering:** GridPool acts as a foundational layer. Traditional pools (PPLNS/FPPS) or larger solo operations can layer their schemes on top of GridPool to "join forces," effectively reducing collective variance while maintaining local block construction rights. #### 7\. Consensus Mechanisms: Dynamic Snapshot Convergence A "Team" is defined as a loose convergence of miners working on templates derived from the same active payout snapshot and sharing proofs into the same unpaid Work Set. To handle the realities of network latency and race conditions, GridPool applies a variant of the "Heaviest Chain" rule: the **Strongest Team Rule**. **Conflict Resolution:** When network latency causes divergent payout snapshots (e.g., if a new high-difficulty share is discovered simultaneously with a Bitcoin block), there could arise a “pool split”, analogous to a chain split in P2Pool like methods. In this case, some miners will be using a snapshot which includes the last minute share, and some will not, depending on whether they observed the new Bitcoin block first or the last minute share first. Competing snapshots and their backing proofs are shared widely across the network. In this event, nodes calculate the total difficulty of each proof-backed snapshot and Work Set by summing the proofs-of-work of the listed shares. The snapshot representing the strongest observed work is selected as the primary list. Nodes naturally converge on this "Strongest Team," resolving splits and joins during the transition between Bitcoin blocks. This also serves as a censorship resistance mechanism. Miners that choose to censor valid shares will fork themselves off to a lower total difficulty set of lists. Game theoretically, miners are incentivized to join the group with the highest total hashpower, as this maximizes their chance of winning a real payout. #### 8\. Algorithmic Scaling: Work Set Reserve and Admission Floor GridPool maintains network performance through a bounded unpaid Work Set reserve, which balances participation accessibility against bandwidth constraints. The reference beta keeps up to three snapshots worth of unpaid proofs: `299 * 3 = 897` proofs by default. A node can advertise the current admission floor for entering that reserve. This lets honest peers avoid relaying shares that cannot affect the payout snapshot, while giving nodes a simple way to identify peers that keep sending obviously too-low-difficulty spam. #### 9\. Sybil Resistance GridPool is fundamentally Sybil-resistant because attribution is tied to verified Proof-of-Work. Identities do not vote; only the realized difficulty of submitted shares determines ranking within the unpaid Work Set and the active payout snapshot.