A Modular Sovereign Stack
Designed for Network Dominance
Sentrix operates as three interconnected modules — each deployable independently in air-gapped or logically isolated VPCs with zero disruption to production traffic.
Intercept
Sentrix-Sentry scores traffic at the edge
Bridge
Diode encrypts one-way telemetry
Attribute
Sentrix-Core generates forensic dossiers
Sentrix-Sentry
The Interceptor
Deployed at the client's network edge, Sentrix-Sentry performs real-time behavioral analysis on every request. Using Kinetic Indicator (KI) scoring, it identifies anomalous patterns — timing jitter, entropy levels, non-human cadence — and makes sub-millisecond decisions.
Stack
Nginx / OpenResty / Lua
Location
Client-Side Edge
Latency Impact
< 2ms added to request
Deploy Mode
Passive → Kinetic toggle
-- Sentrix-Sentry: Kinetic Indicator Scoring
local payload = ngx.req.get_body_data()
local entropy = calculate_entropy(payload)
local jitter = ngx.req.start_time() - ngx.now()
-- AI-lite: High entropy + non-human timing
if entropy > 0.85 or math.abs(jitter) > 0.05 then
ngx.log(ngx.WARN,
"[NVByte-Sentry] KI Detected. "
.. "Signaling Sentrix-Core...")
-- Handshake: Call the Sentrix-Core
local res = http_post(SENTRIX_CORE_API, {
ip = ngx.var.remote_addr,
entropy = entropy,
jitter = jitter,
request_path = ngx.var.uri
})
-- Redirection Decision
if res.status == 302 then
return ngx.redirect(
NEXUS_NODE .. ngx.var.uri)
end
endThe Diode
The Bridge
The Diode establishes a cryptographically enforced one-way channel between the client's Sentry and the NVByte Sentrix-Core. Telemetry flows out — no management access flows in. High-Side tradecraft never leaks to the Low-Side network.
Stack
Encrypted RabbitMQ / Amazon MQ
Encryption
FIPS 140-3 Compliant
Direction
One-Way (Outbound Only)
Protocol
Encrypted Heartbeat Validated
Sentrix-Core
The Brain
The Sentrix-Core is the intelligence engine of Sentrix. Powered by a Hybrid AI model combining behavioral transformers and Graph Neural Networks, it autonomously generates forensic dossiers with actor-level attribution.
Stack
Python FastAPI / Hybrid AI / GNN
Engine
Autonomous Attribution Engine
Confidence
High-Fidelity Actor-Level
Oversight
Human-in-the-Loop (HITL)
# Sentrix-Core: Autonomous Attribution Engine
@app.post("/v1/singularity/score")
async def evaluate_threat(
telemetry: ThreatTelemetry
):
# 1. Run the Transformer Model
ci_score = ai_model.predict(telemetry)
if ci_score > 0.90:
# 2. TRIGGER HITL
await notify_commander(
telemetry, ci_score
)
# Wait for manual authorization
if not commander.is_authorized(
telemetry.id
):
return {
"status": "OBSERVE",
"action": "PASS_THROUGH"
}
# 3. KINETIC ACTION
return JSONResponse(
status_code=302,
content={
"action": "REDIRECT_TO_NEXUS"
}
)Zero-Disruption Deployment
In Five Phases
Reconnaissance
Infrastructure audit, VPC flow log validation, and cross-account IAM provisioning.
Sentry Deployment
Terraform-driven edge deployment in Monitor-Only mode for 24-hour baseline observation.
Diode Handshake
Encrypted MQ tunnel establishment with FIPS 140-3 compliant heartbeat validation.
Nexus Infrastructure
Nexus Node activation with AI-generated honey data and synthetic environments.
Operational Handover
Toggle from PASSIVE to KINETIC engagement. Dashboard access and ROE confirmation.