Session 1.5 - Peer-to-Peer Networks

Learning Objectives

By the end of this session, you will be able to:

Understand routing mechanisms in peer-to-peer networks
Explain how blockchain networks handle node churn
Describe different P2P network topologies and their trade-offs
Analyze network discovery and bootstrap processes
Evaluate the challenges and solutions in P2P network design

What are Peer-to-Peer Networks?

Core Definition

A peer-to-peer (P2P) network is a distributed network where participants (peers) communicate directly with each other without relying on centralized servers or intermediaries.

Traditional Client-Server

Central server controls all data
Clients request information from server
Single point of failure
Server bears all computational load

Peer-to-Peer

No central authority
Peers communicate directly
Distributed and resilient
Load shared among all peers

Real-World Examples

BitTorrent: File sharing without central servers
Bitcoin: Cryptocurrency transactions
Skype (original): Voice calls through peer connections
IPFS: Distributed file storage system

P2P Network Topologies

Unstructured

Random connections between peers

Pros: Simple, robust
Cons: Inefficient routing

Structured

Organized topology (e.g., DHT)

Pros: Efficient routing
Cons: Complex maintenance

Hybrid

Combines structured and unstructured

Pros: Balanced approach
Cons: Increased complexity

Blockchain Networks

Most blockchain networks (like Bitcoin and Ethereum) use unstructured P2P networks because they prioritize robustness and decentralization over routing efficiency.

Routing Mechanisms

Message Propagation Strategies

Flooding

How it works: Each node forwards messages to all its neighbors

Advantages:

Simple to implement
Guarantees message delivery
Highly fault-tolerant

Disadvantages:

High network overhead
Message duplication
Potential network congestion

Random Walk

How it works: Message is forwarded to a randomly selected neighbor

Advantages:

Low network overhead
No message duplication
Scalable approach

Disadvantages:

No delivery guarantee
Potentially slow propagation
May miss some nodes

Bitcoin's Approach

Bitcoin uses a modified flooding approach:

Each node forwards transactions to ~8 peers
Uses exponential backoff to prevent spam
Implements duplicate detection to avoid loops
Prioritizes recent and high-fee transactions

Network Discovery and Bootstrap

The Bootstrap Problem

How does a new node joining the network find other peers to connect to? This is the fundamental bootstrap problem in P2P networks.

Common Bootstrap Methods

Seed Nodes

Hardcoded list of well-known nodes

Pros: Simple, reliable initial connection

Cons: Centralization risk, maintenance overhead

DNS Seeds

DNS queries return IP addresses of active nodes

Pros: Dynamic, can return multiple addresses

Cons: Relies on DNS infrastructure

Peer Exchange

Existing peers share addresses of other peers

Pros: Fully decentralized, self-maintaining

Cons: Requires initial connection

DHT Bootstrap

Distributed Hash Table for peer discovery

Pros: Efficient, structured approach

Cons: Complex implementation

Handling Network Churn

What is Churn?

Churn refers to the continuous process of nodes joining and leaving the network. In blockchain networks, nodes may go offline due to:

Network connectivity issues
Hardware failures
Software updates
Voluntary disconnection

Churn Management Strategies

Keep-Alive Messages

Regular ping/pong messages to detect node failures

Periodic health checks
Timeout-based detection
Connection maintenance

Connection Redundancy

Maintain multiple connections to different peers

Multiple peer connections
Automatic reconnection
Load balancing

Peer Address Caching

Store addresses of known peers for future connections

Local peer database
Address freshness tracking
Reputation scoring

Ethereum's Approach

Ethereum handles churn through:

Kademlia DHT: Structured peer discovery
Node Discovery Protocol: UDP-based peer finding
Connection Limits: Maximum of 25 outbound connections
Peer Scoring: Reputation-based connection management

Security Considerations

Common Attacks

Eclipse Attack: Isolating a node from honest peers
Sybil Attack: Creating multiple fake identities
DDoS: Overwhelming nodes with requests
Routing Attacks: Manipulating message paths

Defense Mechanisms

Diverse Connections: Connect to peers from different subnets
Rate Limiting: Limit requests per peer
Proof of Work: Computational cost for participation
Reputation Systems: Track peer behavior

Eclipse Attack Example

An attacker controls all of a victim's peer connections, effectively isolating them from the honest network. This can lead to:

Double-spending attacks
Censorship of transactions
Mining on outdated chains

Performance Optimization

Key Performance Metrics

Latency

Time for message propagation

Throughput

Messages processed per second

Scalability

Performance with network size

Resilience

Fault tolerance capability

Optimization Techniques

Connection Management: Optimal number of peer connections
Message Prioritization: Important messages first
Compression: Reduce bandwidth usage
Caching: Store frequently accessed data locally
Load Balancing: Distribute traffic across peers

Session Summary

Key Takeaways

P2P networks enable decentralized communication without central servers
Routing mechanisms like flooding ensure message propagation across the network
Network churn is handled through redundancy and peer discovery protocols
Bootstrap methods help new nodes join the network initially
Security threats like eclipse and sybil attacks require specific defenses
Performance optimization balances latency, throughput, and resilience

What's Next?

In the next session, we'll explore Proof of Work, diving into the hash-puzzle mechanism that secures Bitcoin and understanding how mining difficulty adjusts to maintain consistent block times.

Next: Proof of Work Previous: GARAY & RLA Models Back to Course Home