Skip to content

7-POD Stack

7-POD Stack

HookProbe uses a modular POD architecture where each POD handles a specific function. This separation enables scaling, isolation, and maintainability.

POD Overview

PODFunctionKey Services
POD-001Web UI / API Gateway / DMZNginx, API Gateway
POD-002IAM / Identity / RBACKeycloak, SSO
POD-003Persistent DatabasePostgreSQL
POD-004Transient Data / CacheRedis/Valkey
POD-005Metrics / DashboardsPrometheus, Grafana
POD-006Kernel Packet IntakeAegis (Zig + eBPF/XDP)
POD-007AI Detection + ResponseNapse (Mojo AI), AEGIS Orchestrator
POD-008Workflow Automationn8n, Playbooks
POD-009NotificationsEmail, Cloudflare Tunnel
POD-010DSM LedgerConsensus Engine

POD-001: Web Management UI

Purpose: Entry point for all web and API traffic.

Services

ServicePortFunction
Nginx Proxy80, 443TLS termination, routing
API Gateway8080Request routing, rate limiting
Static Assets-Dashboard UI

Configuration

pod-001:
proxy:
ssl_enabled: true
rate_limit: "100r/s"
upstream:
- dashboard: 172.30.0.11:3000
- api: 172.30.0.20:8080

POD-002: Identity & Access

Purpose: Authentication, authorization, and user management.

Services

ServicePortFunction
Keycloak8080OIDC provider
LDAP Sync-Directory integration

Roles

RolePermissions
master_adminFull platform access
adminTenant administration
soc_analystSecurity analysis
editorContent management
customerStandard user

POD-003: Persistent Database

Purpose: Durable data storage.

Services

ServicePortFunction
PostgreSQL5432Primary database
Backup Agent-Automated backups

Schemas

-- Device registry
CREATE TABLE devices (
device_id TEXT PRIMARY KEY,
device_type TEXT NOT NULL,
hardware_fingerprint TEXT NOT NULL,
status TEXT NOT NULL,
first_seen TIMESTAMP,
last_seen TIMESTAMP
);
-- DSM checkpoints
CREATE TABLE dsm_checkpoints (
epoch INTEGER PRIMARY KEY,
merkle_root TEXT NOT NULL,
validator_id TEXT NOT NULL,
timestamp TIMESTAMP,
agg_signature TEXT
);

POD-004: Cache & Queues

Purpose: Fast transient data and message queuing.

Services

ServicePortFunction
Valkey/Redis6379Cache, sessions
Message Queue-Event distribution

Use Cases

Data TypeTTLPurpose
Sessions24hUser authentication
Rate limits1minRequest throttling
Pending intents5minIntent aggregation
Model cache1hML inference cache

POD-005: Metrics & Dashboards

Purpose: Observability and visualization.

Services

ServicePortFunction
Prometheus9090Metrics collection
Grafana3000Dashboards
VictoriaMetrics8428Long-term storage

Key Metrics

MetricTypeDescription
qsecbit_scoreGaugeCurrent security score
aegis_packets_droppedCounterAegis XDP DDoS mitigation
napse_intents_totalCounterNapse intent classifications
dns_blocks_totalCounterdnsXai blocks

POD-006: Kernel Packet Intake (Aegis)

Purpose: Kernel-level packet intake, feature extraction, and early filtering.

Services

ServiceFunction
Aegis XDPZig + eBPF kernel-level packet processing
Ring Buffer ManagerBPF ring buffer for zero-copy Aegis-to-Napse communication

Aegis Pipeline

Packet arrives at NIC
|
v
Aegis XDP (kernel) -- Drop DDoS, rate limit
|
v
Feature extraction (entropy, protocol class, 32-byte vector)
|
v
BPF Ring Buffer (16 MB shared memory, zero-copy)
|
v
Napse consumer (POD-007)

POD-007: AI Detection + Response (Napse + AEGIS)

Purpose: AI-powered intent classification and autonomous threat response.

Napse (Mojo AI) consumes observations from the Aegis ring buffer and performs Bayesian intent classification. AEGIS (Adaptive Endpoint Guardian with Intelligent Security) is HookProbe’s AI orchestrator that bridges all detection systems into unified reasoning with plain-English explanations.

Services

ServiceFunction
Napse (Mojo AI)SIMD-vectorized Bayesian intent classification
Napse HMMKill chain state tracking (Viterbi decoding)
dnsXaiAI DNS protection
Layer DetectorL2-L7 analysis
AEGIS Orchestrator8-agent AI coordination
ML InferenceThreat classification
Response EngineAutomated actions

Specialized Agents (AEGIS)

AgentDomainFunction
GUARDIANNetwork DefenseDDoS, port scans, L3-L4 attacks
WATCHDOGDNS ProtectionDGA detection, DNS tunneling
SHIELDEndpoint ProtectionDevice classification, policy
VIGILAuthorizationTLS downgrade, cert issues
SCOUTReconnaissancePort scans, enumeration
FORGEHardeningProactive audits, configuration
MEDICIncident ResponseCross-layer correlation, forensics
ORACLEForecastingTrend analysis, recommendations

Detection Pipeline

Aegis Ring Buffer (from POD-006)
|
v
Napse Bayesian Classifier -- Intent classification
|
v
Napse HMM -- Kill chain state tracking
|
v
dnsXai -- DNS classification
|
v
QSecBit -- Unified score
|
v
AEGIS -- Autonomous response

Response Actions

TriggerAction
QSecBit AMBERIncrease logging
QSecBit REDEnable full mitigation
High-confidence C2 intentBlock source via Aegis XDP
C2 detectedQuarantine device

POD-008: Workflow Automation

Purpose: Security orchestration.

Services

ServicePortFunction
n8n5678Workflow automation
Playbook Engine-Defensive playbooks

Example Workflows

  • Threat -> Slack notification -> Ticket creation
  • Alert -> Email -> Escalation
  • Detection -> Block -> Forensic capture

POD-009: Notifications

Purpose: External communication.

Services

ServiceFunction
Email ServiceAlert notifications
Cloudflare TunnelSecure external access
Webhook DeliveryIntegration events

POD-010: DSM Ledger

Purpose: Decentralized consensus.

Services

ServiceFunction
DSM NodeMicroblock creation
DSM ValidatorCheckpoint aggregation
Consensus EngineBLS signature aggregation
DSM APIQuery interface

Data Structures

Microblock:

{
"type": "M",
"node_id": "edge-uuid-12345",
"seq": 1847,
"prev": "hash-of-previous",
"timestamp": "2025-12-07T18:35:00Z",
"payload_hash": "sha256-of-event",
"signature": "tpm-signed"
}

Checkpoint:

{
"type": "C",
"epoch": 147,
"merkle_root": "root-hash",
"included_ranges": {
"edge-12345": [1840, 1850]
},
"agg_signature": "bls-aggregated"
}

Network Topology

+---------------------------------------------------------------+
| POD Network (172.30.0.0/24) |
+---------------------------------------------------------------+
| |
| +---------+ +---------+ +---------+ |
| | POD-001 | | POD-002 | | POD-003 | |
| | Proxy | | IAM | | DB | |
| | .2 | | .40 | | .30 | |
| +----+----+ +----+----+ +----+----+ |
| | | | |
| -----+--------------+--------------+---------- |
| | |
| -----+--------------+--------------+---------- |
| | | |
| +----+----+ +---------+ +----+----+ |
| | POD-004 | | POD-005 | | POD-006 | |
| | Cache | | Metrics | | Aegis | |
| | .31 | | .50 | | .60 | |
| +---------+ +---------+ +----+----+ |
| | |
| +----+----+ |
| | POD-007 | |
| | Napse | |
| | .61 | |
| +---------+ |
| |
+---------------------------------------------------------------+

Deployment

Start All PODs

Terminal window
podman-compose up -d

Start Specific POD

Terminal window
podman-compose up -d pod-006

View POD Status

Terminal window
podman ps -a | grep hookprobe

View POD Logs

Terminal window
podman logs hookprobe-<pod-name> --tail 50

Resource Allocation

PODCPURAMStorage
POD-0010.5256MB100MB
POD-0021.0512MB100MB
POD-0031.01GB10GB
POD-0040.5256MB1GB
POD-0051.0512MB5GB
POD-006 (Aegis)1.01GB500MB
POD-007 (Napse + AEGIS)2.02GB1GB

Next Steps