UTMStack v11 introduces a revolutionary architecture designed for modern cybersecurity operations. The platform offers flexible and scalable deployment models that adapt to organizations of any size, from small businesses to large enterprises and MSPs.

New in v11: The architecture now features a manager-worker model with horizontal scalability, replacing the monolithic design of previous versions.


Core Architecture Components

Manager and Worker Nodes

UTMStack v11 uses a distributed architecture with two primary container types:

Manager Node

Central coordination and management

  • Web interface hosting

  • User authentication and authorization

  • Configuration management

  • Alert orchestration

  • API endpoints

  • Database management

Worker Nodes

Distributed data processing

  • Log ingestion and parsing

  • Real-time correlation

  • Threat detection

  • Plugin execution

  • Parallel processing

  • Horizontal scalability

Key Architectural Features

EventProcessor Engine

Replaces Logstash with a custom-built, high-performance log processing engine developed by Threatwinds:

  • Lower resource consumption

  • Faster processing speeds

  • Better memory management

  • Native correlation capabilities

Modular Plugin System

Official plugin architecture for extensibility:

  • Independent feature modules

  • Easy maintenance and updates

  • Community contributions support

  • Hot-swappable components

Horizontal Scaling

Add worker nodes to scale processing capacity:

  • Linear performance scaling

  • No single point of failure

  • Load distribution across workers

  • Automatic failover support


Deployment Models

UTMStack v11 supports multiple deployment models to meet different organizational needs:

1. Single Node Deployment (Standalone)

Single Node Deployment

Best for: Small to medium organizations with up to 500 data sources

A single node deployment combines both manager and worker functions on one server, providing:

  • Complete data control and isolation

  • Simplified management

  • Lower infrastructure costs

  • Suitable for on-premises or cloud deployment

  • Full SIEM/XDR capabilities

Characteristics:

  • All-in-one installation

  • Direct log collection from devices

  • Local data storage

  • No external dependencies

  • Enhanced security through data isolation


2. Multi-Node Deployment (Manager + Workers)

Multi-Node Deployment

Best for: Large organizations with 500+ data sources requiring high performance

A multi-node deployment separates management from processing:

  • One manager node for coordination

  • Multiple worker nodes for data processing

  • Horizontal scalability

  • High availability options

  • Load distribution

Characteristics:

  • Scalable architecture

  • Add workers as needed

  • Parallel log processing

  • Better resource utilization

  • Handles large data volumes efficiently


3. Federated Deployment (MSP Model)

Federated Master Deployment

Best for: Managed Service Providers (MSPs) managing multiple customers

The federated deployment model enables centralized management across multiple UTMStack instances:

  • Separate UTMStack installation per customer

  • Central federation server for unified monitoring

  • Multi-tenant architecture

  • Centralized alerting and reporting

  • Customer data isolation

Characteristics:

  • One instance per customer network

  • Central monitoring dashboard

  • Unified alert management

  • Efficient multi-customer oversight

  • Scalable MSP operations

This model is commonly used by SOC teams for its simplicity and effectiveness in managing multiple client environments.


4. SaaS Deployment (Fully Managed)

SaaS Deployment

Best for: Organizations preferring a fully managed solution

The SaaS deployment model provides a turnkey, cloud-hosted solution:

  • Hosted and managed by UTMStack

  • Automatic updates and scaling

  • High availability included

  • Professional support

  • No infrastructure management

Characteristics:

  • Cloud-based deployment

  • Agents or SyslogTLS for log collection

  • Automatic backups

  • 24/7 monitoring

  • Managed updates and maintenance

  • Focus on your business, not infrastructure


Data Flow Architecture

graph TB
    A[Data Sources] --> B[Collection Layer]
    B --> C[EventProcessor]
    C --> D[Correlation Engine]
    D --> E[Storage Layer]
    D --> F[Alert Engine]
    F --> G[SOAR Workflows]
    E --> H[Query Engine]
    H --> I[Web Interface]
    F --> I
    G --> I

Processing Pipeline

  1. Data Collection: Agents, syslog, APIs collect logs from sources

  2. EventProcessor: Parses and normalizes incoming data

  3. Correlation Engine: Real-time correlation before storage

  4. Storage Layer: Elasticsearch for indexed log storage

  5. Alert Engine: Generates alerts based on correlation rules

  6. SOAR: Automated response workflows

  7. Query Engine: Fast search and analysis

  8. Web Interface: User interaction and visualization


Security Architecture

Encryption in Transit

  • TLS 1.3 for all connections

  • Certificate-based authentication

  • Encrypted agent communication

Data Isolation

  • Container isolation

  • Network segmentation

  • Encrypted data at rest

  • Secure credential storage

Access Control

Mandatory Multi-Factor Authentication

  • Role-based access control (RBAC)

  • Session management

  • Audit logging

security

Service Security

  • Microservices architecture

  • Fail2ban protection

  • Regular security updates

  • Vulnerability scanning


Scalability Considerations

When to Scale Horizontally

Monitor Performance Metrics

Watch CPU, memory, and disk I/O on your manager node

Add Worker Nodes When
  • Processing more than 500 data sources

  • CPU usage consistently above 70%

  • Log ingestion delays occur

  • Real-time correlation lags

Scale Gradually

Add worker nodes one at a time and monitor improvement

Optimize Distribution

Configure plugin distribution across workers for optimal performance


High Availability Options

For mission-critical deployments:

  • Database Clustering: Elasticsearch cluster for data redundancy

  • Manager Redundancy: Active-passive manager configuration

  • Worker Pools: Multiple workers ensure continued processing

  • Load Balancing: Distribute user connections across manager nodes

  • Backup Systems: Automated backup and disaster recovery


Network Architecture

Required Connectivity

Manager Node:
  ← Data Sources (various ports)
  ← Worker Nodes (internal)
  ← Administrators (443/TCP)
  → Central Server (optional)

Worker Nodes:
  ← Data Sources (various ports)
  → Manager Node (internal)
  → Elasticsearch (internal)

Security Zones

  • DMZ: Agent collectors and log receivers

  • Internal: Core processing and storage

  • Management: Web interface and administration

  • Isolated: Customer data in federated deployments


Comparison: v10 vs v11 Architecture

Featurev10v11
Processing EngineLogstashEventProcessor
ScalabilityVertical onlyHorizontal + Vertical
ArchitectureMonolithicDistributed (Manager/Worker)
Plugin SystemIntegratedModular
Resource UsageHigherSignificantly lower
MFAOptionalMandatory
Central ManagementLimitedFull support
Auto UpdatesManualAutomatic (optional)

Choosing Your Deployment Model

Single Node

Choose if:

  • < 500 data sources

  • Budget-conscious

  • Simple management preferred

  • Single location deployment

Federated

Choose if:

  • MSP or MSSP

  • Multiple customers

  • Centralized monitoring needed

  • SOC operations

Multi-Node

Choose if:

  • 500 data sources

  • High performance required

  • Large data volumes

  • Enterprise scale

SaaS

Choose if:

  • No infrastructure team

  • Prefer managed solution

  • Quick deployment needed

  • Focus on operations not maintenance