IBM Maximo Application Suite (MAS): Complete Architecture & Components Guide

The IBM Maximo Application Suite represents the most significant architectural transformation in Maximo’s 40-year history. Moving from a traditional on-premise J2EE application to a containerized, AI-enabled cloud suite changes not just the technology stack but the operational model, the licensing economics, and the pace of innovation that Maximo customers can expect. This guide provides a comprehensive technical overview of MAS for architects, administrators, and technical decision-makers evaluating or planning a MAS deployment.

MAS Architecture: OpenShift as the Foundation

Every MAS application runs as a containerized workload on Red Hat OpenShift Container Platform. This architectural decision has profound implications for how MAS is deployed, operated, and upgraded compared to Maximo 7.6.

Red Hat OpenShift Container Platform (OCP) is a Kubernetes distribution with enterprise features: integrated container registry, software-defined networking, role-based access control, and certified operator lifecycle management. IBM delivers MAS through the OpenShift Operator Framework — each MAS application has an OCP Operator that manages its deployment, configuration, and upgrades.

The IBM MAS Operator orchestrates the installation and configuration of all MAS applications in the cluster. Administrators interact with MAS primarily through the OpenShift web console (for infrastructure-level operations) and the MAS Administration console (for application-level configuration such as user management, workspace configuration, and license allocation).

Core platform dependencies that MAS requires beyond OCP:

• IBM Cloud Pak Foundational Services (CPFS): provides identity management, licensing metering, and common platform services
• IBM Db2 or Oracle Database: MAS Manage requires a supported relational database. Db2 11.5+ or Oracle 19c are the current supported options.
• MongoDB: used by MAS Monitor and MAS Predict for IoT time-series data and machine learning model metadata
• IBM Db2 Warehouse (optional): for advanced analytics workloads in MAS Health and Predict
• Red Hat SSO (Keycloak): identity provider for all MAS applications

The storage requirements for a full MAS deployment are substantial. A production MAS environment with Manage, Monitor, and Health requires approximately 2-5 TB of persistent storage across the applications, depending on IoT data volume and historical data retained.

MAS Manage: The EAM Core

MAS Manage is the cloud-native successor to Maximo 7.6. It preserves the full EAM data model and business process capabilities of Maximo 7.6 while adding:

Modernized UI framework: the MAS Manage UI is built on IBM Carbon Design System, providing a cleaner, more responsive interface compared to the Maximo 7.6 Application Framework. The navigation structure has been updated, though core application names and workflows remain familiar to Maximo 7.6 users.

Maximo Mobile: the rebuilt technician mobile application (described in the work orders guide) runs natively on iOS and Android with offline capabilities, GPS-based work order navigation, and barcode scanning.

Automation Scripts as the customization model: Java class customization is replaced by Automation Scripts, providing a version-stable extensibility model that survives MAS upgrades without recompilation.

Native MAS integration: Manage shares a data layer with Monitor, Health, and Predict. When Monitor detects an anomaly on a monitored asset, it can automatically create a Manage work order using internal APIs — no external integration configuration required.

MAS Monitor: Industrial IoT Platform

MAS Monitor is IBM’s industrial IoT monitoring application for connected assets. It ingests time-series sensor data, applies analytics rules for anomaly detection and threshold alerting, and surfaces equipment performance dashboards.

Data ingestion: Monitor connects to industrial data sources through adapters for OSIsoft PI, MQTT brokers, OPC-UA servers, and cloud IoT platforms (AWS IoT Core, Azure IoT Hub). Sensor data is stored in Monitor’s time-series data store (backed by MongoDB) and can be retained for months or years depending on the subscription tier.

Analytics rules: the Monitor rules engine evaluates sensor data streams against configurable thresholds and anomaly detection algorithms. Rules can trigger: alerts displayed in the Monitor dashboard, email or mobile notifications to maintenance personnel, and automatic work order creation in MAS Manage.

Anomaly detection: beyond simple threshold rules, Monitor applies statistical anomaly detection that learns the normal operating signature of each monitored asset and flags deviations from that signature — catching developing failures that do not yet exceed static threshold limits.

Asset monitoring dashboards: Monitor provides configurable dashboards showing real-time sensor readings, historical trends, and alert history for each monitored asset. These dashboards are visible to both reliability engineers and plant operators, providing a shared operational view of equipment condition.

MAS Health: Asset Condition Scoring

MAS Health addresses a fundamental challenge in maintenance management: how do you make apples-to-apples comparisons of asset condition across a fleet of hundreds or thousands of different equipment types?

Health solves this by calculating a normalized condition score (0-100) for each asset, aggregating multiple data inputs — work order history, meter readings, age versus design life, inspection results, and sensor anomalies from Monitor — into a single comparable number.

Score cards: the configuration layer for Health scores, defining which data inputs contribute to the score and their relative weights. A score card for rotating equipment might weight work order frequency at 40%, operating hours versus design life at 30%, and recent anomaly alerts at 30%.

Fleet view: the Health application’s primary visualization shows all assets in a selected class or location on a condition score ranking — from best condition to worst. Maintenance managers can identify the most vulnerable assets in their fleet at a glance and prioritize inspection or replacement decisions accordingly.

Investment planning: Health includes a capital replacement planning module that forecasts the cost of maintaining versus replacing each low-scoring asset based on its historical maintenance cost trajectory and replacement value. This data feeds directly into capital budget requests with objective, data-driven justification.

For organizations evaluating predictive maintenance capabilities, Health provides the condition scoring foundation that Predict’s failure prediction models build upon.

MAS Predict: AI Failure Prediction

MAS Predict is the AI analytics application that transforms historical maintenance data and real-time sensor readings into forward-looking failure predictions.

Training data: Predict trains machine learning models on work order history from MAS Manage (failure dates, failure codes, maintenance actions) combined with time-series sensor data from Monitor. The training process identifies the sensor signatures and maintenance patterns that precede failure events, creating failure prediction models specific to each asset class in the fleet.

Failure probability scores: for each monitored asset, Predict calculates a failure probability score for defined time horizons (next 7 days, next 30 days, next 90 days). Assets with high near-term failure probability appear prominently in the Predict dashboard, allowing maintenance managers to schedule interventions before the failure occurs.

Remaining Useful Life (RUL): for assets with sufficient historical data, Predict estimates the remaining operating time before the asset reaches end of reliable life. RUL estimates are presented as probability distributions rather than single-point estimates, reflecting the inherent uncertainty in predicting physical system behavior.

Automated work order generation: when a Predict failure probability score crosses a configured threshold, Predict automatically creates a work order in MAS Manage with the predicted failure mode, recommended maintenance action, and supporting evidence from the sensor data — enabling truly autonomous condition-based maintenance without requiring a reliability engineer to review every alert.

MAS Visual Inspection: Computer Vision for Asset Health

MAS Visual Inspection (formerly known as Maximo Visual Inspection or PowerAI Vision) applies computer vision and deep learning to detect defects in images captured during equipment inspections.

The application supports two inspection workflows:

Inspection routes: maintenance technicians capture images of equipment using the MAS Mobile application during rounds. Visual Inspection analyzes each image automatically and classifies it as pass, fail, or flagged — detecting defects such as corrosion, cracks, fluid leaks, or insulation damage that a human inspector might miss on a quick visual check.

Fixed-camera monitoring: cameras mounted on production lines or at fixed equipment locations continuously stream images to Visual Inspection. The AI models detect anomalies in real time — a new crack in a weld, an unexpected particle in a product stream, a displaced component — and trigger alerts or work orders immediately.

Training a Visual Inspection model requires labeling a sufficient number of example images (typically 50-200 per defect class) through the Visual Inspection labeling tool. IBM provides pre-trained models for common industrial inspection scenarios, reducing the data collection burden for standard use cases.

AppPoints Licensing: Understanding Consumption

The AppPoints model is IBM’s consumption-based licensing approach for MAS. Organizations purchase a pool of AppPoints and allocate them across applications and user types.

Manage AppPoints consumption: Maximo Manage users consume different AppPoints quantities depending on their usage level — approximately 60 points per full user (equivalent to a “premium” Maximo 7.6 license), 30 points per “operator” user (equivalent to “base”), and 10 points per “limited” user.

Add-on application consumption: Monitor, Health, Predict, Visual Inspection, and Assist each consume additional AppPoints on top of the base Manage allocation. The consumption rate scales with usage volume (number of monitored assets for Monitor, number of predicted assets for Predict) rather than user count.

AppPoints flexibility: organizations can reallocate AppPoints between applications as their usage evolves. If Monitor adoption grows faster than Predict, AppPoints can be shifted accordingly without repurchasing licenses. This flexibility is a genuine advantage over the Maximo 7.6 model, where each module required separate license purchases.

IBM provides AppPoints consumption tracking through the MAS Administration console, allowing organizations to monitor their allocation utilization and project future consumption as new applications or user groups are added.

Update Cadence and Lifecycle Management

MAS follows a continuous delivery model. IBM releases monthly cumulative updates for each MAS version, containing security patches, bug fixes, and new features. Administrators apply updates through the MAS Operator, which orchestrates a rolling update across all application pods without requiring application downtime in properly configured high-availability deployments.

This monthly update cadence is a significant operational change from Maximo 7.6, where quarterly fix packs required careful testing and change management processes. Organizations migrating to MAS must establish a regular patching process and allocate dedicated time for testing each monthly update in a non-production environment before applying it to production.

IBM’s lifecycle policy for MAS versions provides approximately 24-36 months of support for each major version before requiring an upgrade to the next major version. This is a shorter lifecycle than Maximo 7.6 (which has been in extended support since 2023 and will receive support through at least 2027), reflecting the faster pace of cloud software evolution.

For organizations with strict change control requirements — regulated industries, critical infrastructure operators — the monthly update cadence requires developing a lightweight but rigorous testing process that can be executed within the change control window without blocking critical security patches. The Industrie du Futur platform tracks how Industry 4.0 technology stacks, including cloud EAM suites like MAS, are evolving to meet enterprise operational technology requirements.