Security & Trust

Built For
Critical
Infrastructure.

Industrial operations run 24/7. A breach isn't just a data problem - it's an operational one. Every layer of Merobix is designed with security first, from the gateway at your wellhead to the infrastructure that stores your historian data - in our cloud, or on your own air-gapped servers.

TLS 1.2+
All data in transit; 1.3 preferred
MFA
Available on all accounts
99.9%
Uptime Target
RLS
Database-level tenant isolation
Dozens
Of tables under row-level security
SOC 2-Aligned Controls C1D2-Rated Hardware Options

Three Pillars of
Merobix Security

Protect
Connections to the platform encrypted with TLS. Passwords stored only as salted hashes. Sessions behind authenticated sign-in, with MFA available on all accounts - from sensor to screen.
Isolate
Your data is yours alone. Postgres Row-Level Security across dozens of tables, tenant-scoped MQTT topics, and per-device credentials. Isolation is enforced by the database itself - not just application code.
Audit
Every login, every config change, every alert acknowledgment is logged with timestamp, user, and IP. Tamper-evident records available to you at any time.

Defense In Depth,
Layer By Layer

No security features locked behind enterprise tiers - every Merobix customer gets the full stack from day one. Security is built into the platform at seven layers: identity, tenant isolation, device and telemetry, control safety, the network edge, data protection, and detection.

Identity & Access Protection

Multi-Factor Authentication
TOTP authenticator apps plus FIDO2/WebAuthn hardware security keys and passkeys - phishing-resistant MFA, with MFA seeds encrypted at rest and lockout after repeated failed codes.
Six Permission Roles + Site Scoping
Viewer, Operator, Manager, Engineer, Administrator, and Superadministrator roles checked on every API endpoint - with site-level authorization restricting users to their assigned facilities.
Hardened Sessions
256-bit random session tokens with idle expiry, concurrent-session limits, instant revocation, and logout-everywhere - a stolen laptop doesn’t stay logged in.
Brute-Force Protection
Persistent per-account and per-IP failure counters stored in the database - lockouts survive restarts and autoscaling. Account lockout plus separate MFA-attempt lockout.
Step-Up Authentication
High-risk workflows - such as permit signing - require re-verifying your identity in the moment, even inside an already-authenticated session.
Scoped API Keys
API keys carry explicit scopes with full lifecycle controls - creation, expiry, and immediate revocation - so an exposed key never grants more than it was issued for.
Enterprise SSO
SAML 2.0, OIDC (with strict token and JWKS validation), RADIUS, and PKI / X.509 client-certificate authentication for companies with existing identity infrastructure.
Password Policy & Hashing
Passwords hashed with PBKDF2-HMAC-SHA256 at 200,000 iterations with per-user salt and constant-time verification, plus enforced complexity and reuse prevention. Plaintext is never stored or logged.

Customer & Tenant Isolation

Postgres Row-Level Security
Tenant isolation enforced inside the database itself - RLS policies on dozens of tables are designed so one customer’s query returns only that customer’s rows, even if application code is bypassed.
Signed Database Request Context
The tenant context every query runs under is cryptographically signed - application code cannot be tricked into switching to another customer’s context.
Separate Service & Tenant Contexts
Background services and customer requests run under distinct database contexts, so platform plumbing never shares an identity with tenant data access.
Cross-Tenant Denial Alerts
Every denied cross-tenant request is logged and raises a security alert - probing between customers is detected, not silently dropped.
Site Ownership Checks
Operational and administrative actions verify site ownership before executing - an authenticated user still can’t act on a facility they don’t own.

Device & Telemetry Security

Unique Device Identity
Every field device registers individually with its own credentials - plus time-limited, single-use enrollment tokens that are consumed on first use and cannot be replayed.
Ed25519 Device Signing Keys
Gateways generate their own Ed25519 signing identities - the private key is created on the device and proves which physical device produced each reading.
Signed Telemetry Envelopes
MQTT telemetry travels in signed envelopes carrying sequence numbers, timestamps, sample identities, and payload digests - with replay detection rejecting recorded-and-resent data.
Durable Ingestion & Store-and-Forward
Telemetry is durably accepted before historian processing, and gateways buffer locally through cellular or broker outages - no silent data loss, no gaps an attacker can hide in.
TLS in Transit
TLS for cloud, MQTT, and industrial protocol connections that support it - TLS 1.2 minimum, TLS 1.3 preferred, weak ciphers prohibited outright.
Signed Gateway Updates
Gateway software artifacts are signed and verified before installation - a tampered update simply will not run.
Per-Device MQTT Credentials
Each field device gets its own broker credentials and a topic ACL scoped to your company - a compromised device credential cannot read or publish into anyone else’s data stream.

Operational Control Safety

Explicit Writable-Tag Allowlist
Nothing is writable by default. Control is only possible on tags explicitly configured as writable - with control-tag authorization and setpoint bounds checked before dispatch.
Role & Competency Checks
Control actions verify both the user’s role and their competency records - and management-of-change and pre-startup safety-review gates apply where configured.
Idempotent Commands with States
Duplicate command protection plus explicit pending, applied, verified, failed, and expired states - you always know exactly what a command did.
Read-Back Verification
For protocols that support it, the platform reads the value back from the device after a write to confirm the setpoint actually landed.
Command Audit Records
Every operational command and acknowledgement is recorded - who, what, when, and the outcome.

Application & Network Edge

Cloudflare Edge Protection
The cloud edge sits behind Cloudflare with HTTPS enforcement, HSTS, a restrictive Content Security Policy, clickjacking protection, and MIME-sniffing protection.
Locked-Down Production Surface
Production API documentation disabled, sensitive file paths blocked, trusted-origin CORS with no wildcards, and coarse public health responses that don’t leak internal infrastructure details.
Rate Limits & Request Caps
Per-endpoint rate limits, request-size limits, strict input validation, and bounded outbound requests on every route.
SSRF & Webhook Hardening
Outbound connectors and webhooks carry SSRF defenses - including pinned-IP webhook delivery with original-host TLS validation, so a webhook URL can’t be turned against internal systems.
SQL Injection Protection
Every database query is parameterized, read-only SQL paths are validated as read-only, and dynamic identifiers are restricted to strict allowlists.

Data Protection & Integrity

Encryption of Sensitive Data
Sensitive application data - MFA seeds, phone numbers, push tokens - is encrypted at the application layer with key-rotation support, on top of provider disk encryption.
Blind-Index Search
Encrypted identifiers like email addresses are searchable through blind indexes - lookups work without ever storing the plaintext value.
Hashed One-Time Tokens
Password-reset and registration tokens are stored hashed - a database leak doesn’t hand out working reset links.
Tamper-Evident Audit Chain
Sensitive events are logged in immutable, HMAC-chained records - any tampering with historical entries is cryptographically detectable.
Full Telemetry Provenance
Every sample carries its identity end to end - sample ID, timestamp, quality, device, site, and tenant - so operational evidence stands up to scrutiny.
Validated Backups & Export
Backup archives are validated, restores go through controlled tooling, and data-retention and customer-export workflows keep your data portable and yours.

Detection, Alerting & Auditability

Runtime Threat Monitoring
Live detection of denied cross-tenant requests, privilege probes, repeated authentication failures, and suspicious control activity.
Durable Security Notifications
Security events flow through a durable outbox with retry and dead-letter handling - delivered by email, SMS, webhook, or straight into your SIEM.
SIEM Integration
Security events stream in real time via Syslog and standard connectors to your existing SIEM for centralized monitoring and alerting.
Health & Freshness Surfaces
Dedicated health endpoints for API dependencies and fleet telemetry freshness - you can see the platform’s own vital signs.
Audit-Chain Verification
Administrators can verify the audit chain’s integrity on demand - proving records haven’t been altered since they were written.
Build & SHA Visibility
Deployed services expose their build and Git SHA - you can always tie the running platform to an exact, reviewable release.

What We're
Building Next

Security is never finished. These hardening programs are in active development right now - we publish them here because a vendor that only shows you the finished list is hiding the work in progress.

Complete Telemetry Enforcement
Closing remaining fallback-path gaps so signed telemetry is mandatory across MQTT and HTTP store-forward delivery - with a controlled signing-identity migration for legacy gateways.
Safer Gateway Updates
Immutable versioned install directories, atomic activation and rollback, removal of stale executables, artifact provenance bound to the exact cloud release, and hardware validation on physical RUT956 gateways.
Stronger Control-Command Outcomes
Explicit driver write outcomes - NOT_SENT, REJECTED, SENT_UNCONFIRMED, CONFIRMED - distinguishing safe retries from ambiguous physical writes, preventing duplicate actuation and stranded commands.
Release Assurance
Blocking build, test, SAST, dependency, secret, and contract checks; deployment only from an exact green SHA; immutable release bundles; signed build provenance; and a sealed release-control environment.
Database Cutover Hardening
Mandatory runtime context keys, complete privilege preflight before any migration, atomic baseline creation, verified least-privilege runtime identities, and repeatable rollback drills.
Alert Delivery Assurance
Provider canaries proving at least one alert channel works before readiness reports healthy, sticky degraded status when a security event can’t be persisted or delivered, and duplicate-alert reconciliation.
Continuous Independent Validation
Authenticated cross-tenant testing, gateway-to-cloud replay and forgery testing, dependency and static-analysis scanning, restore and failover exercises, regular security testing, and SOC 2 readiness with evidence collection.

Our Security
Position

No platform is "unhackable," and we will never tell you Merobix is. Our goal is to reduce attack paths, limit the impact of compromised credentials or devices, detect abnormal activity quickly, and preserve trustworthy operational evidence.

We are not yet SOC 2 certified, and we won't claim to be until an independent audit is complete - our SOC 2 readiness program, regular security testing, and continuous validation work (authenticated cross-tenant testing, gateway replay and forgery testing, restore and failover exercises) are part of the ongoing program above.

Security at Merobix is an ongoing engineering and operational discipline, not a one-time feature. If you want the details, ask - we'll walk your team through the architecture directly, and you can read how we think about industrial security in our security engineering guides - 60+ articles on SCADA security, compliance, and certifications.

Outbound-Only.
No Open Ports.

The Merobix gateway initiates all connections outbound. Your OT network never accepts inbound connections from the internet - no firewall rules to write, no VPN to configure, no attack surface exposed.

Your PLC / Sensor
Modbus · OPC-UA · DNP3
OT NETWORK (isolated)
Serial / Ethernet
Merobix Gateway
C1D2 options · Local buffer
No inbound ports open
TLS 1.2+ outbound
Merobix Cloud API
Railway · FastAPI · Auth
HTTPS · MFA session
Your Browser
scada.merobix.com
MFA verified · RBAC enforced

Every Protocol.
Secured.

Merobix ships drivers across the major protocol families. Each one is handled with appropriate security controls - encrypted where the protocol supports it, authenticated at the gateway level where it doesn't.

Protocol Transport Security handling Common use
Modbus RTUSerial RS-485 Gateway-level isolation. Serial data never reaches the internet - encrypted at gateway before cloud push. Wellhead controllers, RTUs, flow computers
Modbus TCPEthernet TCP/IP Local LAN only. Gateway reads via private IP. No Modbus traffic exposed to internet. PLCs, VFDs, process equipment
OPC-UAEncrypted TCP/IP Native OPC-UA security modes supported - Sign, SignAndEncrypt. Certificate-based authentication. Siemens and Rockwell PLCs
MQTTTLSAuthenticated TCP/IP Private HiveMQ Cloud broker with TLS (1.2 minimum, 1.3 preferred) and per-customer credentials. No public broker exposure. IoT sensors, edge devices, 4G gateways
DNP3Serial / TCP Serial / TCP Gateway-isolated. DNP3 Secure Authentication v5 supported where device firmware allows. Water utilities, substations, SCADA RTUs
BACnet/IPEthernet UDP/IP Local network only. Firewall rules isolate BACnet traffic to building LAN. No internet exposure. HVAC, building automation
EtherNet/IPEthernet TCP/IP Private LAN only. Gateway connects to CIP devices on isolated OT network segment. Allen-Bradley PLCs, Rockwell systems
HTTP POSTHTTPSToken auth HTTPS Bearer token authentication required. HTTPS enforced - plain HTTP rejected. Rate limited per device. Custom sensors, data loggers, edge computers

Common
Questions

Is my sensor data encrypted in transit?

+

Yes. All data between your gateway and the Merobix cloud is encrypted with TLS - 1.2 minimum, 1.3 preferred, the same standard used by major banks. Data is never transmitted in plaintext at any point in the pipeline, from sensor reading to browser dashboard.

Can other Merobix customers see my data?

+

No. Isolation is enforced at the database layer with PostgreSQL Row-Level Security on dozens of tables - the policies are designed so one customer's query returns only that customer's rows, even if application code were bypassed. Historian data and MQTT topics are also scoped per customer with per-device credentials and topic ACLs.

Do I need to open firewall ports or set up a VPN?

+

No. The Merobix gateway initiates all connections outbound over HTTPS or cellular. No inbound ports need to be opened, no VPN tunnels required, and no changes to your site's firewall. Your OT network stays completely isolated from the internet.

What authentication does Merobix use?

+

Every user account supports multi-factor authentication - TOTP authenticator apps or FIDO2/WebAuthn hardware keys and passkeys. Sessions use 256-bit random tokens, expire after 8 hours of inactivity, and are capped at 3 concurrent per user. Failed logins trigger persistent per-account and per-IP lockouts. Passwords are hashed with PBKDF2-HMAC-SHA256 at 200,000 iterations - never stored in plaintext.

Who can access my account and make changes?

+

Access is controlled by role - Viewer, Operator, Engineer, and Admin, plus custom per-company roles and per-site scoping. Each role has strict permission boundaries enforced at the API level. Viewers and operators can work with live data but cannot change configurations. Only admins can add users or modify alert thresholds. Logins, configuration changes, and alert acknowledgments are logged in an HMAC-chained, tamper-evident audit trail.

What happens if the gateway loses internet connectivity?

+

The gateway buffers data locally during any outage and automatically syncs when connectivity is restored. You will receive an offline alert via SMS and email, typically within a minute of the gateway going dark. Your PLC and local control logic continue to operate independently of the cloud - Merobix is a monitoring layer, not a control dependency.

Is Modbus secure? It doesn't have built-in encryption.

+

Modbus RTU and TCP don't have native encryption - that's a known limitation of the protocol. Merobix handles this by keeping all Modbus traffic on your private local network or serial bus. The gateway reads locally and encrypts data before it ever leaves your site. The Modbus protocol itself never touches the internet.

How are new user accounts created?

+

Users register with their company email domain. New registrations are held in a pending state and require explicit approval from your company admin or a Merobix superadmin before access is granted. No self-service access to live operational data - every account is verified before activation.

Can Merobix staff access my data?

+

Merobix superadmin accounts are restricted to the management portal only - they cannot log into your SCADA dashboard or view your operational data. Access to customer data for support purposes requires documented justification and is always logged in the audit trail.

What happens if I suspect a breach?

+

Contact [email protected] immediately. We can revoke all active sessions for your account, force password resets for all users, and provide a full audit log export on request. We take security incidents seriously and respond promptly during business hours.

Is Merobix SOC 2 certified?

+

Not yet - and we won't claim to be until an independent audit is complete. Merobix runs a SOC 2 readiness program with evidence collection, maps its controls to IEC 62443 and NIST SP 800-82, and runs regular security testing in its ongoing validation program. We publish exactly where we are rather than overstating certifications, and we're happy to share our current control documentation with your security team under NDA.

TLS 1.2+ Encrypted
MFA Available on All Accounts
availability per the applicable agreement
Full Audit Log
Role-Based Access
C1D2-Rated Hardware Options

Questions About
Our Security?

Talk to us directly. We'll walk you through the architecture, answer your compliance questions, and show you exactly how your data is protected.