Protocol Overview
Centrix operates as a distributed protocol with no central authority. The network consists of independent participants who coordinate through smart contracts, peer-to-peer communication, and cryptographic mechanisms.Network Participants
Requestors (Demand Side) Submit computational tasks to the network, specify resource requirements, set maximum acceptable prices, and verify results. Providers (Supply Side) Register hardware with the network, set pricing and availability schedules, bid on tasks, execute tasks in isolated environments, and submit results for verification. Validators Operate verification nodes, validate computation results, maintain consensus on task outcomes, earn fees for services, and can be slashed for dishonest behavior. Token Holders Provide liquidity and stability, earn staking rewards from protocol fees, and participate in governance decisions.Task Lifecycle
A typical task progresses through the following states:- Automatic progression for successful tasks
- Timeout mechanisms prevent stalling
- Dispute resolution for contested results
- Checkpoint recovery for interrupted tasks
Economic Primitives
CXT Token- Native currency of the Centrix network
- Required for all task payments
- Staked by validators and premium providers
- ERC-20 compatible for broad accessibility
- Dynamic pricing based on supply and demand
- Providers set base rates + bid adjustments
- Requestors set maximum acceptable price
- Market clearing through competitive bidding
- Premium pricing for urgent tasks
Protocol Basics
Task Specification
Tasks are defined using a standardized specification format:- Requirements: Hardware specifications needed
- Input: Data sources (IPFS, HTTP, S3-compatible)
- Container: Docker image and execution command
- Verification: How results will be validated
- Payment: Pricing and escrow details
Provider Discovery
Providers advertise their capabilities through a distributed hash table (DHT): Provider Registration:- Filter providers meeting minimum requirements
- Rank by combination of price and reputation
- Request bids from top N candidates
- Requestor selects based on bid competitiveness
Execution Environment
Tasks execute in containerized environments providing: Security Isolation:- Separate namespaces (PID, network, mount)
- Resource limits (cgroups)
- Read-only root filesystem
- Dropped dangerous capabilities
- seccomp/AppArmor profiles
- Real-time CPU/memory/network tracking
- Alert system for limit violations
- Automatic throttling if exceeded
- Usage data for billing accuracy
- Regular checkpointing of progress
- State saved to persistent storage
- Restart from last checkpoint on failure
- Automatic provider reassignment if unresponsive
Result Verification
Centrix employs multiple verification strategies depending on task criticality: 1. Redundant Execution- Task executed by 3+ independent providers
- Results compared using cryptographic hashes
- Consensus determined by majority agreement
- Dissenters penalized through reputation slashing
- Provider generates ZK proof of correct execution
- Validators verify proof without re-execution
- Dramatically reduces verification cost
- Applicable to certain computational classes
- Random sampling of task results
- Re-execution on trusted validator nodes
- Statistical guarantee of correctness
- Cost-effective for large batches
- Providers stake CXT proportional to task value
- Stake slashed if fraud detected
- Makes dishonesty economically irrational
- Bonds returned after successful verification
Architecture
System Components
Layer 1: Blockchain Layer- Ethereum mainnet for settlements and token
- Smart contracts for payments and escrow
- Layer 2 (Optimistic Rollups) for scalability
- Cross-chain bridges for multi-chain support
- Task queue and distribution
- Provider discovery (DHT)
- Bid collection and matching
- State management and consensus
- Containerized compute environments
- Resource isolation and monitoring
- Result collection and submission
- Local storage and caching
- Result validation nodes
- Consensus mechanisms
- Dispute resolution
- Reputation tracking
- IPFS for distributed storage
- BitTorrent for large file distribution
- CDN integration for performance
- Encrypted data transmission
Network Topology
Communication Protocols
Peer-to-Peer (P2P)- libp2p protocol stack
- Gossipsub for message propagation
- DHT (Kademlia) for peer discovery
- NAT traversal (hole punching, TURN relays)
- RESTful HTTP API for requestors
- WebSocket for real-time updates
- gRPC for high-performance internal communication
- GraphQL for complex queries
- IPFS for content-addressed storage
- BitTorrent for redundant downloads
- Direct peer-to-peer when possible
- Fallback to CDN for reliability
Smart Contract Architecture
Core Contracts:- TaskManager.sol - Task lifecycle, bid management, escrow
- PaymentProcessor.sol - Token transfers, fee distribution, slashing
- ReputationRegistry.sol - Reputation scores, performance data, history
- VerificationArbitration.sol - Dispute resolution, validator voting, appeals
Scalability Architecture
Horizontal Scaling- Add more providers → Linear capacity increase
- Sharded task queues for parallelism
- Geographic distribution reduces latency
- Optimistic Rollups for 10,000+ TPS
- Batch processing for gas efficiency
- Off-chain computation, on-chain settlement
- Provider metadata cached locally
- Popular task templates pre-loaded
- IPFS data pinned on multiple nodes
- CDN for static resources
- Task Submission: < 1 second
- Provider Matching: < 5 seconds
- Result Verification: < 60 seconds
- Payment Settlement: < 2 minutes
- Network Capacity: 1M+ concurrent tasks

