internetcomputer

END-TO-END PROJECT: BUILD A DECENTRALIZED TODO LIST
Let us build a complete application from scratch.
PROJECT SETUP:
1. Install dfx
2. Create project: dfx new todo-app
3. Choose Motoko or Rust
4. Configure dfx.json
BACKEND CANISTER:
actor TodoApp {
stable var todos: [(Nat, Text, Bool)] = [];
stable var nextId: Nat = 0;
public func addTodo(text: Text) : async Nat {
let id = nextId;
todos := Array.append(todos, [(id, text, false)]);
nextId += 1;
return id;
};
public func completeTodo(id: Nat) : async () {
todos := Array.map(todos, func((tid, text, done)) {
if (tid == id) (tid, text, true)
else (tid, text, done)
});
};
public query func getTodos() : async [(Nat, Text, Bool)] {
return todos;
};
}
FRONTEND:
Create index.html with:
1. Input field for new todos
2. Add button
3. List of todos
4. Complete button for each
DEPLOYMENT:
1. dfx start --background
2. dfx deploy
3. Open in browser
4. Test functionality
PRODUCTION CHECKLIST:
1. Add authentication (Internet Identity)
2. Add stable memory
3. Add error handling
4. Add frontend certification
5. Deploy to mainnet
EXTENSIONS:
1. Add user accounts
2. Add categories
3. Add due dates
4. Add reminders
5. Add sharing
THIS IS THE COMPLETE ICP COURSE.
From introduction to architecture to building real applications. You now have the knowledge to build anything on the Internet Computer.
The future is decentralized. The future is ICP.
$ICP #InternetComputer #CourseComplete

ARCHITECTURE CASE STUDIES — REAL-WORLD ICP APPLICATIONS
CASE STUDY 1: OPENCHAT
Architecture:
- Messaging canister: Handles messages
- User canister: Manages profiles
- Group canister: Group chat logic
- Frontend canister: UI serving
Key decisions:
- Separate canisters for different functions
- Stable memory for message history
- Internet Identity for authentication
- SNS for decentralized governance
CASE STUDY 2: SONIC DEX
Architecture:
- Swap canister: Core trading logic
- Pool canister: Liquidity management
- Factory canister: Creates new pools
- Frontend canister: Trading interface
Key decisions:
- High-performance Rust canisters
- Composite queries for price data
- ckBTC for Bitcoin liquidity
- Multiple canisters for parallel processing
CASE STUDY 3: ICPSWAP
Architecture:
- Order book canister: Limit orders
- AMM canister: Automated market making
- Governance canister: Protocol upgrades
- Analytics canister: Market data
Key decisions:
- Hybrid order book + AMM model
- Cross-canister calls for complex trades
- Stable memory for order history
- SNS governance for decentralization
LESSONS LEARNED:
1. Start simple, add complexity gradually
2. Separate concerns into different canisters
3. Use stable memory for critical data
4. Plan for upgrades from day one
5. Monitor everything
This is how real production applications are designed on ICP.
Next: Part 36 — Complete End-to-End Project
$ICP #InternetComputer #Course

ICP ECOSYSTEM — WHAT EXISTS TODAY AND WHAT IS COMING
CURRENT ECOSYSTEM (2026):
1. DEFI
- ICPSwap: Leading DEX
- Sonic: AMM protocol
- Hot or Not: Social-fi platform
- ICP.Lending: Lending protocol
2. SOCIAL
- OpenChat: Messaging platform
- DSCVR: Social media
- Taggr: Blogging platform
- Hot or Not: Content rating
3. TOOLS
- Plug Wallet: Browser wallet
- Internet Identity: Authentication
- NFID: Identity provider
- Stöpsi: Hardware wallet
4. INFRASTRUCTURE
- CycleDAO: Cycle management
- ICP.rocks: Analytics
- ICP Dashboard: Network monitoring
- Canistergeek: Developer tools
FUTURE ROADMAP:
1. CHAIN FUSION
- Connect to more blockchains
- Solana integration
- Cosmos integration
- Polkadot integration
2. SCALING
- More subnets
- Higher throughput
- Lower latency
- Better developer tools
3. AI INTEGRATION
- On-chain AI models
- AI-powered dApps
- Decentralized AI training
- AI governance
4. ENTERPRISE
- Enterprise adoption
- Compliance frameworks
- Enterprise tooling
- Industry-specific solutions
THE BIG PICTURE:
ICP is building the infrastructure for a decentralized internet. Every piece — from cryptography to governance to tokenomics — works together to create a platform that can replace Big Tech.
This is not just another blockchain. This is the future of the internet.
Next: Part 34 — Interview Questions
$ICP #InternetComputer #Course

MONITORING AND UPGRADES — KEEPING YOUR DAPP ALIVE
PRODUCTION MONITORING:
1. Canister metrics
- Instruction consumption
- Memory usage
- Cycle balance
- Message throughput
2. Alerting
- Low cycle balance
- High error rates
- Performance degradation
- Security events
3. Logging
- Structured logging
- Event tracking
- Error capture
- Audit trails
TOOLS:
1. dfx canister status
2. IC Dashboard
3. Canister snapshots
4. Custom monitoring canisters
UPGRADE STRATEGIES:
1. Blue-green deployment
- Deploy new version alongside old
- Test new version
- Switch traffic
- Keep old version as backup
2. Rolling upgrade
- Upgrade canisters one by one
- Minimal downtime
- Gradual rollout
3. Hot upgrade
- Upgrade without stopping
- Use stable memory
- Preserve all state
UPGRADE CHECKLIST:
1. Backup state
2. Test upgrade on testnet
3. Verify state preservation
4. Check controller access
5. Monitor after upgrade
ROLLBACK PLAN:
1. Keep previous version ready
2. Test rollback procedure
3. Document rollback steps
4. Set up automated rollback
This is how you maintain production-grade applications on ICP.
Next: Part 33 — Ecosystem & Future Roadmap
$ICP #InternetComputer #Course

BUILDING A PRODUCTION DAPP ON ICP
Step-by-step guide to building and deploying a production-ready application.
PHASE 1: PLANNING
1. Define the problem
2. Choose your architecture
3. Select the right language
4. Design the data model
5. Plan for upgrades
PHASE 2: DEVELOPMENT
1. Set up dfx locally
2. Create canister project
3. Write backend canisters
4. Write frontend code
5. Test locally
PHASE 3: TESTING
1. Unit tests
2. Integration tests
3. Load testing
4. Security audit
5. Testnet deployment
PHASE 4: DEPLOYMENT
1. Deploy to mainnet
2. Configure controllers
3. Set up monitoring
4. Fund canisters with cycles
5. Launch to users
PHASE 5: MAINTENANCE
1. Monitor performance
2. Respond to issues
3. Release updates
4. Scale as needed
5. Community management
BEST PRACTICES:
1. Start with a minimum viable product
2. Iterate based on user feedback
3. Design for failure
4. Implement proper error handling
5. Document everything
COMMON MISTAKES:
1. Overengineering from the start
2. Ignoring security
3. Not planning for upgrades
4. Poor error messages
5. Inadequate testing
This is how professional ICP applications are built.
Next: Part 32 — Monitoring & Upgrades
$ICP #InternetComputer #Course

DEFI ON ICP — DECENTRALIZED FINANCE INFRASTRUCTURE
ICP provides the foundation for building sophisticated DeFi applications.
CORE COMPONENTS:
1. Tokens
- ICP native token
- ckBTC and ckETH
- SNS tokens
- Custom tokens via ledger canister
2. DEX Infrastructure
- Order book DEX
- AMM (Automated Market Maker)
- Liquidity pools
- Cross-chain swaps
3. Lending Protocols
- Collateralized lending
- Flash loans
- Yield farming
- Interest rate models
ADVANTAGES OF ICP FOR DEFI:
1. Low fees (fractions of a cent)
2. Fast finality (200ms)
3. No gas volatility
4. Canister-to-canister calls
5. Bitcoin and Ethereum integration
KEY PROTOCOLS:
1. Sonic DEX: Automated market maker
2. ICPSwap: Order book and AMM
3. Hot or Not: Decentralized content with economic model
4. OpenChat: Social platform with payment integration
CROSS-CHAIN DEFI:
1. Bitcoin DeFi via ckBTC
2. Ethereum DeFi via ckETH
3. Unified liquidity across chains
4. No bridge risk
BUILDING BLOCKS:
1. Ledger canister for token management
2. DEX canister for trading
3. Oracle canister for price feeds
4. Governance canister for protocol upgrades
This is the future of finance: open, transparent, and accessible to everyone.
Next: Part 30 — ckBTC, ckETH and Chain Fusion
$ICP #InternetComputer #Course

SCALING — FROM 100 TO 100 MILLION USERS
ICP scales horizontally by adding subnets. Here is how to build for massive scale.
HORIZONTAL SCALING:
1. Multiple canisters for different functions
2. Geographic distribution of canisters
3. Load balancing across subnets
4. Automatic scaling through NNS
CANISTER ARCHITECTURE:
1. Microservices approach
2. Each function in its own canister
3. Inter-canister communication
4. Independent scaling per canister
DATA PARTITIONING:
1. Shard data across multiple canisters
2. Use stable memory for large datasets
3. Implement data pagination
4. Cache frequently accessed data
CACHING STRATEGIES:
1. Query calls are fast (no consensus)
2. Use composite queries for complex reads
3. Cache results in heap memory
4. Invalidate cache strategically
LOAD TESTING:
1. Simulate concurrent users
2. Measure response times under load
3. Identify bottlenecks
4. Optimize hot paths
MONITORING:
1. Track canister performance
2. Monitor subnet utilization
3. Set up alerts for degradation
4. Plan capacity ahead
PRODUCTION CHECKLIST:
1. Multiple canisters for redundancy
2. Geographic distribution
3. Automatic failover
4. Performance monitoring
5. Cost optimization
This is how ICP scales to support the next billion users.
Next: Part 29 — DeFi Infrastructure
$ICP #InternetComputer #Course

AI ON ICP — THE CONVERGENCE OF AI AND BLOCKCHAIN
ICP is positioning itself as the platform for AI-powered decentralized applications.
AI INTEGRATION OPTIONS:
1. HTTPS Outcalls
Canisters call external AI APIs (OpenAI, Anthropic, etc.)
- Send prompts to AI models
- Process responses on-chain
- Store results in stable memory
2. On-Chain AI Models
Run small AI models directly in canisters:
- Decision trees
- Simple neural networks
- Rule-based systems
- Limited by compute constraints
3. AI Oracles
Specialized canisters that provide AI services to other canisters:
- Natural language processing
- Image classification
- Predictive analytics
- Sentiment analysis
USE CASES:
1. AI-Powered DeFi
- Automated trading strategies
- Risk assessment
- Portfolio optimization
2. AI Governance
- Proposal analysis
- Voting recommendations
- DAO automation
3. AI Content
- Content generation
- Translation
- Summarization
4. AI Security
- Fraud detection
- Anomaly detection
- Smart contract auditing
CHALLENGES:
- Compute limitations on-chain
- Latency for real-time AI
- Cost of running AI models
- Data availability for training
FUTURE VISION:
As ICP scales and AI models become more efficient, fully on-chain AI agents will become possible. This could revolutionize how we think about decentralized intelligence.
Next: Part 26 — Performance Optimization
$ICP #InternetComputer #Course

AZLE — TYPESCRIPT ON ICP
Azle is a TypeScript/JavaScript framework for building ICP canisters. It brings the familiar JavaScript ecosystem to blockchain development.
WHY AZLE:
1. Write canisters in TypeScript
2. Use npm packages
3. Familiar JavaScript tooling
4. Full Candid type generation
5. React-like frontend integration
BASIC EXAMPLE:
import { query, update, StableBTreeMap } from azle;
let counter: StableBTreeMap<string, number> = new StableBTreeMap(counter, 0, 100, 1000);
#[update]
function increment(): void {
counter.insert(count, (counter.get(count) ?? 0n) + 1n);
}
#[query]
function getCount(): bigint {
return counter.get(count) ?? 0n;
}
KEY FEATURES:
1. StableBTreeMap: Persistent key-value storage
2. ManagementCanister: Direct access to IC management
3. HttpTypes: HTTP outcall support
4. Candid: Automatic type generation
5. Test framework: azle test
WHEN TO USE AZLE:
- Full-stack TypeScript developers
- Projects needing npm ecosystem
- Rapid prototyping
- Teams with JavaScript backgrounds
- Projects requiring complex frontend-backend integration
AZLE vs MOTOKO vs RUST:
Azle is easiest for JavaScript developers. Motoko is best for ICP-native development. Rust offers the most performance and control.
This makes ICP accessible to the millions of JavaScript developers worldwide.
Next: Part 24 — Frontend Development
$ICP #InternetComputer #Course

RUST ON ICP — PERFORMANCE AND SAFETY
Rust is a first-class language on ICP. It is used for performance-critical canisters and complex applications.
WHY RUST:
1. Memory safety without garbage collection
2. Zero-cost abstractions
3. Concurrency without data races
4. Excellent performance
5. Growing ICP ecosystem
KEY CRATES:
- ic-cdk: Core CDK for canister development
- ic-cdk-macros: Procedural macros
- ic-stable-memory: Stable memory library
- candid: Candid serialization
BASIC EXAMPLE:
use ic_cdk_macros::query;
#[query]
fn greet(name: String) -> String {
format!("Hello, {}!", name)
}
KEY CONCEPTS:
1. Canister macros: #[update], #[query], #[heartbeat]
2. Call management: ic_cdk::api::call
3. Memory management: ic_cdk::api::stable
4. Time: ic_cdk::api::time
5. Randomness: ic_cdk::api::raw_rand
WHEN TO USE RUST:
- High-performance DeFi protocols
- Complex cryptographic operations
- Memory-intensive applications
- Need for low-level control
- Production systems requiring maximum reliability
RUST vs MOTOKO:
Rust has a steeper learning curve but offers more control and performance. For most projects, Motoko is sufficient. For demanding applications, Rust is the choice.
This is the language for serious ICP development.
Next: Part 23 — Azle TypeScript Development
$ICP #InternetComputer #Course

MOTOKO — THE NATIVE LANGUAGE OF ICP
Motoko is the programming language designed specifically for writing ICP canisters.
WHY MOTOKO:
1. Purpose-built for ICP
2. Native support for canister features
3. Strong type system
4. Easy to learn for JavaScript developers
5. First-class support for actors and messages
BASIC SYNTAX:
actor Hello {
public query func greet(name: Text) : async Text {
return "Hello, " # name # "!";
};
};
KEY CONCEPTS:
1. Actors
Canisters are actors. They communicate via messages. Each actor has its own state.
2. Asynchronous Execution
All inter-canister calls are asynchronous. No blocking.
3. Pattern Matching
Powerful pattern matching for data types. Similar to Rust or Haskell.
4. Stability
The stable keyword preserves state across upgrades.
5. Modules
Code is organized into modules. Import and export functionality cleanly.
MOTOKO vs RUST:
Motoko is easier to learn. Rust is faster and more flexible. Choose Motoko for rapid development. Choose Rust for performance-critical applications.
MOTOKO ECOSYSTEM:
- Vessel: Package manager
- moc: Compiler
- dfx: Build tool
- Motoko Playground: Online IDE
This is where most ICP developers start. Simple, clean, and powerful.
Next: Part 22 — Rust Canister Development
$ICP #InternetComputer #Course

ETHEREUM INTEGRATION — CHAIN FUSION ON ICP
ICP can natively interact with Ethereum. Canisters can read ETH state, sign transactions, and deploy smart contracts.
HOW IT WORKS:
1. Ethereum Adapter
Similar to Bitcoin, each node runs an Ethereum adapter. It connects to Ethereum nodes and reads blockchain data.
2. Threshold ECDSA
ICP nodes collectively sign Ethereum transactions. The same threshold cryptography used for Bitcoin applies here.
3. EVM RPC Canister
A system canister that provides direct access to Ethereum RPC endpoints. Canisters can query ETH balances, call smart contracts, and submit transactions.
WHAT THIS ENABLES:
- Hold ETH on ICP
- Interact with Ethereum smart contracts
- Build cross-chain DeFi applications
- Access the entire Ethereum ecosystem
ckETH:
Chain Key Ethereum (ckETH) is a token on ICP backed 1:1 by real ETH. Similar to ckBTC, it enables:
- Near-instant ETH transfers
- Low fees
- Seamless DeFi integration
- No bridge risk
CROSS-CHAIN USE CASES:
1. Ethereum DeFi on ICP
2. Cross-chain lending and borrowing
3. Multi-chain asset management
4. Ethereum NFT trading on ICP
5. Unified liquidity pools
This is what DFINITY calls CHAIN FUSION — the ability to connect multiple blockchains natively without bridges. Bitcoin, Ethereum, and ICP working together seamlessly.
Next: Part 17 — Threshold ECDSA and Schnorr
$ICP #InternetComputer #Course

BITCOIN INTEGRATION — NATIVE BTC ON ICP
ICP can natively create, sign, and submit Bitcoin transactions. No bridges, no wrapped tokens, no intermediaries.
HOW IT WORKS:
1. Threshold ECDSA
ICP nodes collectively hold BTC private key shares. They sign transactions using threshold cryptography. No single node ever has the full key.
2. Bitcoin Adapter
Each node runs a Bitcoin adapter that connects to the Bitcoin network. It reads blocks, submits transactions, and tracks confirmations.
3. Bitcoin Testnet and Mainnet
ICP connects to both Bitcoin mainnet and testnet. Canisters can hold real BTC.
WHAT THIS ENABLES:
- Create Bitcoin wallets on ICP
- Send and receive BTC from canisters
- Build Bitcoin DeFi applications
- Execute complex Bitcoin scripts
ckBTC:
Chain Key Bitcoin (ckBTC) is a token on ICP backed 1:1 by real BTC. It is created by a canister that holds BTC reserves.
ckBTC advantages:
- Near-instant transfers
- Low fees (fractions of a cent)
- No waiting for Bitcoin confirmations
- Works seamlessly with ICP DeFi
USE CASES:
1. Bitcoin wallets with smart contract features
2. Bitcoin DEXs and AMMs
3. Bitcoin lending protocols
4. Bitcoin NFTs
5. Cross-chain DeFi
This eliminates the need for risky bridges. Your BTC stays on Bitcoin, and ICP signs the transactions.
Next: Part 16 — Ethereum Integration
$ICP #InternetComputer #Course

HTTPS OUTCALLS — CONNECTING TO THE OUTSIDE WORLD
ICP canisters can make HTTP requests to external servers. This is called HTTPS Outcalls. It breaks the traditional blockchain isolation.
HOW IT WORKS:
1. Canister makes an HTTP request (GET, POST, HEAD)
2. All nodes in the subnet independently execute the request
3. Responses are compared using consensus
4. A single agreed-upon response is returned to the canister
WHAT THIS ENABLES:
- Fetch real-time price data from exchanges
- Call traditional APIs (Twitter, GitHub, weather)
- Integrate with Web2 services
- Access off-chain data without oracles
CONSENSUS ON RESPONSES:
For deterministic results:
- GET requests must return the same response
- POST requests must produce the same result
- Headers are normalized
- Timing is managed by the subnet
USE CASES:
1. DeFi: Real-time price feeds
2. Social: Posting to Twitter from a canister
3. IoT: Reading sensor data
4. Enterprise: Calling legacy APIs
5. AI: Fetching external model results
LIMITATIONS:
- Cannot maintain stateful HTTP connections
- Response size has limits
- Cannot use cookies or sessions
- Rate limiting depends on external servers
This is what makes ICP the only fully on-chain platform that can seamlessly interact with the entire web.
Next: Part 15 — Bitcoin Integration
$ICP #InternetComputer #Course

SNS — DECENTRALIZED GOVERNANCE FOR DAPPS
The Service Nervous System allows any canister-based application to become governed by a DAO. It is the framework for decentralized governance on ICP.
WHAT IS AN SNS?
An SNS is a DAO framework. It lets:
- Developers decentralize control of their dApp
- Users participate in governance
- Token holders vote on proposals
- Treasury is managed collectively
SNS LAUNCH PROCESS:
1. Developer submits a proposal to the NNS
2. NNS approves the SNS launch
3. Community swaps ICP for SNS tokens
4. SNS governance is activated
5. Developer transfers control to the SNS DAO
SNS COMPONENTS:
- SNS Governance Canister: Handles proposals and voting
- SNS Ledger Canister: Manages SNS token balances
- SNS Swap Canister: Manages the initial token swap
- SNS Root Canister: Coordinates all SNS canisters
ADVANTAGES:
1. No central authority controls the dApp
2. Community decides on upgrades and features
3. Treasury is governed by token holders
4. Transparent decision-making
EXAMPLES OF SNS DAOS:
- OpenChat: Decentralized messaging
- Hot or Not: Decentralized content rating
- DSCVR: Decentralized social media
- Taggr: Decentralized blogging
Next: Part 13 — Internet Identity
$ICP #InternetComputer #Course

ICP TOKEN — UTILITY AND GOVERNANCE
The ICP token serves three purposes:
1. GOVERNANCE
ICP holders can stake tokens in the NNS to vote on proposals. Voting power is proportional to stake. Staking duration multiplies your voting power (8 year staking = 3x multiplier).
2. CYCLES CONVERSION
ICP can be converted to cycles to power canisters. This is a one-way conversion: ICP goes in, cycles come out. This creates demand for ICP.
3. REWARDS
Node providers are rewarded in ICP. stakers receive voting rewards. This incentivizes participation.
TOKENOMICS:
Total supply: ~520 million ICP
Circulating: ~480 million ICP
Locked in staking: ~300 million ICP
DISTRIBUTION:
- DFINITY Foundation: 23.86%
- Team and advisors: 18.00%
- Public sale: 4.95%
- Seed contributors: 9.50%
- Strategic contributors: 9.93%
- Early token holders: 3.56%
INFLATION AND DEFLATION:
ICP has both inflation and deflation:
- Inflation: Node provider rewards, staking rewards
- Deflation: Cycles burned, governance proposals
The net effect depends on usage. More canisters running = more cycles burned = deflationary pressure.
This creates a direct link between network usage and token value.
Next: Part 11 — Network Nervous System (NNS)
$ICP #InternetComputer #Course

SUBNETS — HOW ICP SCALES HORIZONTALLY
A subnet is a collection of nodes running in consensus. Each subnet is essentially an independent blockchain that hosts canisters.
SUBNET TYPES:
1. System Subnets
Run critical infrastructure like the NNS, CMC, and II. Not for public canisters.
2. Application Subnets
Host user canisters. These are the workhorses of ICP. Each application subnet runs 13-40 nodes.
3. Verified Subnets
Special subnets where all nodes are independently verified by third parties.
HOW REPLICATION WORKS:
Every canister on a subnet is replicated across ALL nodes in that subnet. If a subnet has 13 nodes, your canister runs on 13 machines.
If a node fails, the subnet continues operating. The failed node is replaced automatically by the NNS.
SCALING STRATEGY:
ICP scales by creating MORE subnets, not by making existing ones bigger. This is horizontal scaling. When demand grows, the NNS proposes creating new subnets and migrating canisters.
CANISTER MIGRATION:
Canisters can be moved between subnets without losing state. The NNS coordinates this process. Users never notice the migration.
Next: Part 6 — Canisters Deep Dive
$ICP #InternetComputer #Course

ICP CONSENSUS — HOW AGREEMENT IS REACHED
ICP uses a novel consensus protocol called IC-Consensus. It combines several innovations:
1. THRESHOLD BLS SIGNATURES
Each block proposal is signed by threshold BLS. If enough nodes agree, the block is finalized. This is faster than traditional Nakamoto consensus.
2. ASYNCHRONOUS VERIFIABLE RANDOM FUNCTION (AVRF)
Randomly selects block makers each round. Prevents targeted attacks on specific nodes. Ensures fairness.
3. PIVOTING
If a block maker is slow or malicious, another node takes over. The network dynamically switches to the fastest honest proposer.
4. DAG-BASED CONSENSUS
Instead of a simple chain, ICP uses a directed acyclic graph. Multiple blocks can be proposed in parallel. This increases throughput dramatically.
5. INTER-SUBNET ROUTING
The NNS coordinates messages between subnets. When canister A on subnet 1 calls canister B on subnet 2, the message is routed through a dedicated protocol.
FINALITY:
ICP achieves FINALITY in two steps:
- Verification: A block is verified once enough threshold signatures are collected
- Finalization: A block is finalized when it reaches a certain number of confirmations
This is different from Bitcoin where finality requires 6 confirmations. ICP finality is near-instant.
Next: Part 5 — Subnets and Replication
$ICP #InternetComputer #Course

ICP ARCHITECTURE — HOW IT WORKS INTERNALLY
The Internet Computer is composed of four main layers:
LAYER 1 — HARDWARE (Nodes)
Physical machines running the ICP protocol. Each node runs the IC Replica software. Nodes are organized into subnets.
LAYER 2 — SUBNETS
Groups of nodes that run in consensus. Each subnet is an independent blockchain that canister logic runs on. Subnets are interconnected via NNS-coordinated routing.
LAYER 3 — CANISTERS
Smart contracts on ICP. Unlike Ethereum smart contracts, canisters:
- Have their own memory (heap + stable)
- Can hold web frontend code
- Can be upgraded without losing state
- Execute instructions in deterministic time
LAYER 4 — USERS AND CLIENTS
End users interact with canisters through:
- Internet Identity (authentication)
- Candid (interface description language)
- HTTP outcalls (external APIs)
MESSAGE FLOW:
User -> Gateway -> Subnet Router -> Consensus -> Execution -> Canister -> Response
Each layer handles a specific concern. This modular architecture is what enables ICP to scale horizontally by adding new subnets.
Next: Part 3 — Chain-Key Cryptography
$ICP #InternetComputer #Course

Did you know $ICP constantly processes over 1,000 transactions per second?
That makes the Internet Computer one of the highest-throughput blockchain networks in Web3.
🔹 Real-Time TPS: 1,089 tx/s
🔹 Max TPS (100 blocks): 25,621 tx/s
🔹 Max Theoretical TPS: 209,708 tx/s
🔹 Transaction Volume (1D): 94M txns
🔹 Block Time: 480ms
🔹 Finality: 0 seconds
🔹 Nakamoto Coefficient: 14
🔹 Validators: 673
🔹 Average Tx Fee: $0.00008485
🔹 Total Transactions: 294B
Internet-scale applications need internet-scale performance. $ICP delivers. 🔥
#ICP #InternetComputer #Web3