Infrastructure Overview
Nexis Appchain is a production-grade Layer 3 blockchain built on the OP Stack, specifically optimized for AI agent coordination and verifiable inference workloads. This document provides a comprehensive overview of the technical architecture and design decisions.Architecture Summary
Layer 3 on Base
Built as an OP Stack L3 on top of Base Sepolia (L2), inheriting Ethereum security
2-Second Blocks
High-throughput block production for sub-second transaction finality
Fault Proofs
Permissionless validation with 73-step bisection game for security
EVM Compatible
Full Ethereum compatibility - use existing tools and contracts
Network Specifications
| Parameter | Value | Notes |
|---|---|---|
| Chain ID | 84532 | Unique identifier for the network |
| Block Time | 2 seconds | Consistent block production rate |
| Gas Limit | 30,000,000 | Per-block gas limit |
| Base Fee | 1 gwei | Minimum gas price (EIP-1559) |
| L1 Settlement | Base Sepolia | L2 that batches our L3 transactions |
| Consensus | OP Stack Derivation | Deterministic block derivation from L1 data |
| Finality | ~3-5 minutes | Until L2 batch is confirmed |
OP Stack Layer Architecture
Layer Responsibilities
L1 (Ethereum Mainnet)- Ultimate source of truth and security
- Stores Base L2 state commitments
- Hosts Base L2 fraud proof contracts
- Settles Nexis L3 transactions
- Stores L3 batched transaction data
- Provides data availability guarantees
- Runs fault proof system for L3
- Executes AI agent transactions
- Produces blocks every 2 seconds
- Optimized for high-frequency operations
- Custom smart contracts for agent coordination
Core Components
1. op-geth (Execution Layer)
The execution layer is a modified version of go-ethereum (geth) with OP Stack enhancements:- EVM Execution: Processes smart contract calls and state transitions
- Transaction Pool: Manages pending transactions from users and agents
- State Database: Stores account balances, contract storage, and code
- JSON-RPC: Exposes Ethereum-compatible APIs for wallets and dApps
- Archive Mode: Optional full historical state for analytics
2. op-node (Consensus Layer)
The consensus layer derives L3 blocks from L2 data and manages the rollup protocol:- Block Derivation: Reconstructs L3 blocks from L2 data
- State Synchronization: Keeps execution layer in sync
- P2P Networking: Propagates blocks to other nodes
- Rollup Protocol: Implements OP Stack state transition logic
3. Sequencer
The sequencer is the privileged node that orders and batches transactions: Sequencer Workflow:- Receive transactions from users via JSON-RPC
- Order transactions into a canonical sequence
- Execute transactions through op-geth
- Produce blocks every 2 seconds
- Broadcast blocks to all nodes via P2P
- Batch transactions for L2 submission
Sequencer Operation
4. Batcher
The batcher aggregates L3 transactions and submits them to Base L2 for data availability:- Channel Duration: 30 L2 blocks (~1 minute)
- Max Batch Size: 128 KB per transaction
- Compression: Zlib for data efficiency
- Submission Frequency: Every ~60 seconds or when batch size reached
5. Proposer
The proposer submits L3 state root commitments to Base L2 for finality:- Frequency: Every 120 L3 blocks (~4 minutes)
- Bond Requirement: 1 ETH on Base Sepolia
- Challenge Period: 7 days for fault proofs
- Finalization: After challenge period with no successful disputes
Data Flow
Transaction Lifecycle
State Derivation
Any node can independently derive the canonical L3 chain from L2 data:Security Model
Trust Assumptions
- Liveness: Sequencer must be available (centralized currently, decentralizing soon)
- Data Availability: Base L2 must store and serve batch data
- Fault Proofs: At least one honest verifier must challenge invalid proposals
- L1 Security: Ethereum mainnet remains secure and censorship-resistant
Security Guarantees
- State Validity: Enforced by fault proofs (anyone can challenge invalid state)
- Data Availability: Guaranteed by Base L2 (inherits from Ethereum)
- Censorship Resistance: Users can force inclusion via L2 contracts
- Finality: Economic finality after L2 confirmation, absolute after L1 finalization
Fault Proof System
Nexis uses the OP Stack fault proof game for security:Performance Characteristics
Throughput
| Metric | Value | Notes |
|---|---|---|
| Blocks/second | 0.5 | One block every 2 seconds |
| Gas/block | 30M | Same as Ethereum mainnet |
| Theoretical TPS | ~1,500 | Simple transfers (21,000 gas each) |
| Realistic TPS | ~200-400 | Complex smart contract interactions |
| AI Agent TPS | ~50-100 | Proof submissions (higher gas) |
Latency
| Operation | Latency | Notes |
|---|---|---|
| Transaction inclusion | < 2 seconds | Next block |
| Soft confirmation | ~2 seconds | Sequencer broadcast |
| Safe confirmation | ~4 minutes | After L2 batch + margin |
| Finalized confirmation | ~15 minutes | After L1 (Base) finalization |
| Withdrawal delay | 7 days | Challenge period |
Cost Analysis
Network Topology
Comparison to Other Chains
| Feature | Nexis L3 | Base L2 | Ethereum L1 | Polygon PoS |
|---|---|---|---|---|
| Block Time | 2s | 2s | 12s | 2s |
| Finality | ~15min | ~15min | ~15min | ~30s |
| Gas Cost | 1 gwei | ~0.05 gwei | ~30 gwei | ~100 gwei |
| Security | Fault proofs | Fault proofs | PoS consensus | PoS consensus |
| EVM Compatible | ✅ | ✅ | ✅ | ✅ |
| DA Layer | Base L2 | Ethereum L1 | Ethereum L1 | Ethereum L1 |
Roadmap
Learn More
Consensus Mechanism
Deep dive into block derivation and sequencing
Fault Proofs
Understand the security model and dispute resolution
Run a Node
Participate in the network as a validator
RPC Infrastructure
Learn about the RPC architecture and endpoints
Want to contribute to infrastructure? We’re looking for node operators, validator runners, and infrastructure developers. Join our Discord #infrastructure channel.