WoT Mesh Protocol — Simulation Model

Node State Variables

Per Node

Per Pact

Success Scenarios

Baseline Healthy Network

• Mixed light/full ratio at 75/25
• Random activity distribution across tiers
• High online ratios (>80%)
• WoT graphs with natural social clustering
• Pacts form within 1–2 hops reliably
• Content retrieval latency stays low
• PoS challenges pass consistently
• Renegotiation triggers resolve cleanly without data loss

Power User as Hub

• Full node with high WoT degree
• Serves as bootstrap anchor for new users
• Handles above-average pact load without degrading
• Verify: does karma accumulate proportionally to contribution?

New User Onboarding

• Zero WoT, zero pacts
• Reads relay-curated content
• Builds first 5 mutual follows
• First pact forms — measure time to first pact
• Gradually mesh-native — measure at what WoT size relay dependency drops

Graceful Exit

• Node announces departure
• Handoff period: pact partners find replacements
• Verify: no data loss during transition window
• Measure: how long does full handoff take at different WoT sizes?

Multi-Device User

• Same keypair on mobile + desktop
• Desktop goes offline for 2 weeks
• Mobile maintains pacts independently
• Desktop rejoins — measure resync time and data gap

Rolling Window Expiry

• Events older than 15 days evicted cleanly
• No orphaned storage
• Verify: event retrieval fails gracefully for expired content, doesn't crash

Activity Tier Upgrade

• User goes from 10 events/day to 80 events/day
• Protocol detects delta exceeds renegotiation threshold
• Finds new compatible pact partner within WoT
• Old pact terminates cleanly
• Verify: no storage gap during transition

Attack Vectors

Identity / Sybil Attacks

Basic Sybil

• Attacker creates N fake keypairs
• Attempts to mutual-follow real users to enter WoT
• Measure: how many real follows needed before Sybil enters 1-hop WoT of target?
• Defense variable: WoT score threshold for pact eligibility

Sybil Pact Farm

• Attacker controls 1000 low-activity nodes
• All mutual-follow each other
• Form pacts exclusively within Sybil cluster
• Never actually store data
• Pass PoS challenges by caching temporarily at challenge time
• Measure: does this degrade real network or stay isolated?

Identity Grinding

• Attacker generates keypairs until one has high WoT proximity to target
• Probabilistic — measure expected keypairs needed
• Defense: WoT is built on organic follows, not key proximity

Storage Attacks

Lazy Storage

• Node accepts pact, stores data initially
• Gradually evicts partner data to free space
• Only reconstructs data temporarily when challenge arrives
• Key question: how frequently must challenges occur to make this economically unviable?
• Simulate challenge frequency vs reconstruction cost tradeoff

Storage Inflation

• Node claims high storage capacity to attract pacts
• Actually has low capacity
• Accepts more pacts than it can honor
• Measure: how many pacts does it take before failure cascades?

Selective Storage Breach

• Node stores data for high-karma partners
• Silently drops data for low-karma partners
• Detect via: cross-referencing PoS results across pact network

Replay Attack on PoS

• Node caches a valid challenge response
• Deletes actual data
• Replays cached response to future challenges
• Defense: challenge must include timestamp + nonce, response must be fresh

Eclipse Attack on Storage

• Attacker surrounds target node with Sybil pact partners
• All Sybil nodes fail simultaneously
• Target loses all redundant copies
• Measure: minimum pact count N to survive M simultaneous failures
• This defines your minimum pact redundancy requirement

Network Topology Attacks

WoT Poisoning

• Attacker builds legitimate reputation over time (months)
• Becomes high-WoT node
• Suddenly goes malicious — drops all stored data, rejects challenges
• Long-game attack — measure damage radius at different WoT centrality levels
• Defense: karma decay, pact redundancy, breach propagation speed

Relay Capture

• Attacker controls majority of relays
• Can't affect WoT mesh directly
• Can corrupt discovery layer — surface malicious nodes to new users
• Measure: at what relay capture % does onboarding become compromised?
• Defense: first pacts always within existing WoT, relay only for discovery

Partition Attack

• Network splits into two disconnected WoT clusters
• Relay bridge goes down simultaneously
• Measure: recovery time when bridge relay comes back
• Measure: what % of content is inaccessible during partition?

Graph Fragmentation

• High-degree WoT nodes go offline simultaneously
• Simulate cascading pact failures
• Measure: what's the critical hub removal % before network fragments?
• Directly tests your 75/25 light/full ratio assumption

Karma Attacks

Karma Laundering

• Two colluding nodes exchange fake high-volume pacts
• Inflate each other's karma scores
• Spend inflated karma on real network resources
• Defense: witness nodes must corroborate pact activity
• Measure: collusion size needed to fool witness pool

Karma Sinkhole

• High-karma node accepts pacts, provides good service
• Accumulates massive karma
• Suddenly consumes all karma by requesting maximal resources
• Leaves network without contributing back
• Measure: damage radius, recovery time

Free Rider at Scale

• 20% of network are pure light nodes that never become full nodes
• Never contribute storage
• Only consume via relay and WoT reads
• Measure: at what free rider % does full node load become unsustainable?
• Directly tests whether 75/25 ratio is stable or drifts

Timing Attacks

Challenge Timing Manipulation

• Malicious node monitors challenge intervals
• Reconstructs data just before expected challenge window
• Discards immediately after
• Defense: randomize challenge timing, unpredictable intervals

Renegotiation Flooding

• Attacker triggers mass renegotiation events simultaneously
• Floods pact negotiation layer with proposals
• Legitimate nodes can't find partners during the storm
• Measure: recovery time, % of nodes left without pacts

Churn Storm

• 30% of nodes go offline simultaneously (simulates mobile users sleeping)
• Measure: how many pacts break?
• Measure: how long to reform stable pact graph?
• Measure: what % of content is temporarily unavailable?

WoT-Tiered Read Strategy

The simulator models content retrieval across three trust tiers, reflecting how real social network users consume content from different social distances.

WoT Tier Definitions

2-hop trust is weighted: an author followed by 8 of my mutual contacts is more trusted than one followed by 1.

Read Selection

Each read selects a target author from a tier based on configurable weights (defaults: 50% direct, 30% 1-hop, 20% 2-hop). Per-reader auto-redistribute adapts these weights when a tier is empty or undersized — the budget shifts proportionally to remaining tiers.

Within the 2-hop tier, authors are selected weighted by their trust score (number of shared direct WoT connections).

Trust-Weighted Gossip Routing

Gossip fanout prioritizes high-trust peers. When forwarding a RequestData message, peers are scored by relevance to the target author (direct WoT contact of author → score 3, 2-hop connection → shared count, unknown → 1). Top-scored peers are selected first.

The should_forward() decision also extends to 2-hop trusted peers — messages from senders with known trust scores are forwarded, not just direct WoT contacts.

Expected Behavior

This validates the core thesis: relay dependency drops as social proximity increases.

Validation Tables

Results from 5,000-node, 30-day simulation (BA m=50, seed 42, WoT-tiered reads):

Read Tier by WoT Distance:

Direct WoT reads achieve 98.3% instant delivery, confirming that pact partners reliably store each other's data. 1-hop (non-mutual follows) still achieves 91.9% instant via incidental pact overlap. 2-hop tier is empty at this scale because the BA graph model generates few mutual follow pairs, limiting 2-hop candidate discovery.

Relay Dependency Decay (by reader pact age):

Relay usage drops from 16.9% (pre-pact) to 0.2% (14+ day pact age). This validates the core thesis: relay dependency decays as nodes form and mature pact partnerships.

Content Availability (by simulation period):

Availability improves from 94.5% in the first 5 days to 99.9% by day 20-30, demonstrating the network's self-healing capacity as pacts form and stabilize.

Emergent Behaviors to Watch

Tier Stratification — does the network naturally stratify into stable activity tiers or do tiers blur? Blurring means activity-matching becomes unreliable.

Pact Graph Topology — does the pact graph mirror the WoT graph or diverge? Divergence means activity-matching is overriding social trust — could be good or bad depending on degree.

Relay Dependency Decay Curve — as users age in the network, does relay usage actually drop? Or do users remain relay-dependent indefinitely? This validates or invalidates the core architectural claim.

Karma Distribution — does karma concentrate in power users (Pareto) or distribute evenly? Heavy concentration means the system has a de facto aristocracy of nodes.

Cold Start Equilibrium — what's the minimum viable network size for the mesh to function without relay dependency? Below that threshold the protocol isn't viable standalone.

Simulation Parameters to Sweep

Success Metrics

Priority Simulation

The scenario that matters most: eclipse attack combined with churn storm — simultaneous pact partner failures during a high-churn event. This is the realistic worst case for mobile-heavy networks. If the protocol survives that with acceptable content availability, the architecture is sound.