Whitepaper · DIME for IBM Maximo
Feeding MaximoLive Machine Data
Closing the loop from sensor to work order: the data pipeline that lets IBM Maximo Monitor, Health, and Predict act on real equipment condition.
- The gapWhy condition data stalls
- The pipelineSensor to Maximo with DIME
- The loopAlert to work order to learning
proven. reliable. trusted.
01Executive summary
Maximo can act on machine data. Most of it never arrives.
IBM Maximo Monitor, Health, and Predict are built to turn equipment condition into maintenance decisions. In most plants, the condition data they need is still trapped on the floor, behind dozens of protocols and a separate project for every connection.
Our earlier paper, Addressing Barriers to Advanced Analytics and Predictive Maintenance, described why condition-based and predictive programs stall: equipment diversity, legacy systems, integration complexity, and pilots that never scale. This paper is the follow-on. It describes the data pipeline that removes those barriers for organizations that run, or are moving to, IBM Maximo Application Suite (MAS).
DIME (Data In Motion Enterprise) is an industrial data platform that connects directly to machines and controllers through more than 40 protocol connectors. It gives raw tags the asset context Maximo expects, learns each asset’s normal behavior with continuous anomaly detection, and delivers the result to Maximo. SMS delivers DIME together with its Maximo practice, so the data layer and the system of record are designed as one program.
Many MAS customers already own the capacity to run these applications. AppPoint pools are often sized to headcount rather than actual use, and the headroom can fund Monitor, Health, and Predict for the people who would use them. Section 04 shows how to check, and what DIME adds that AppPoints do not.
This paper is written for the maintenance, reliability, and IT/OT leaders who own Maximo and are expected to show results from it. It covers what each Maximo application needs from the floor, how the pipeline works stage by stage, how to close the loop from anomaly to work order without creating alarm fatigue, and how to start.
- 01
The last mile is the gap
Maximo is ready to act on condition data. Getting it out of PLCs, CNCs, and meters, with asset context, is where programs stall.
- 02
Fewer, better alerts
Anomaly detection that learns each machine’s normal keeps planners focused on deviations that deserve their time.
- 03
Every closed work order teaches
Failure codes and actuals recorded at closure become the history Health and Predict learn from.
02The challenge
Why condition data stalls between the machine and Maximo
The barriers are familiar. What is less obvious is that they all sit in the same place: the layer between the equipment and the system of record.
82%
of asset failures appear random, so calendar-based maintenance alone cannot prevent them.
99%
of data collected from sensor-enabled assets goes unused.
3
Maximo applications, Monitor, Health, and Predict, whose value depends on live condition data.
Failure and sensor-data figures as cited by IBM in its Maximo Manage client presentation.
Every machine speaks its own language
One plant can run Siemens, Allen-Bradley, Fanuc, and Modbus devices side by side. Each needs its own driver, and many older machines have no modern interface at all.
Tags have no asset context
A PLC reports a register address. Maximo expects an asset number, a site, and a meter. Someone has to map one to the other, and keep that map current.
Fixed thresholds generate noise
Static alarm limits fire on every startup, changeover, and seasonal swing. Planners learn to ignore them, and the real warning gets lost with the rest.
Every integration is a project
Custom middleware and point-to-point scripts can carry a pilot line. They rarely survive the second plant, the next upgrade, or the departure of the engineer who wrote them.
03What Maximo needs
What Monitor, Health, and Predict actually consume
Each Maximo application uses condition data differently. A pipeline that serves all of them has to deliver the right shape of data, at the right rate, tied to the right asset record.
| Maximo application | What it needs from the floor | What it returns |
|---|---|---|
| 01Manage | Meter readings and condition-monitoring measurements, recorded against the correct asset or location. | Meter-based PMs, and work orders generated when a measurement crosses its action limit |
| 02Monitor | Streaming device metrics published to the MAS IoT layer, typically over MQTT, with consistent device types. | Dashboards, anomaly functions, and alerts that can be raised in Manage |
| 03Health | Current meter and sensor readings, alongside work order and failure history in Manage. | Health, risk, and criticality scores that rank where to act first |
| 04Predict | Time-series sensor history aligned with recorded failure dates and failure codes. | Failure probability and predicted failure dates for each asset |
| 05YouBIM for Maximo | Live sensor values and alarms keyed to the same assets as Maximo. | Sensor data, alarms, and open work orders in a 3D digital twin of the facility |
04Your AppPoints
You may already own the Maximo side
AppPoints are a shared pool. The same points that cover your Manage users can cover the people who would use Monitor, Health, and Predict, and many pools have room to spare.
28%
median amount by which MAS AppPoint pools exceeded measured use in one licensing analysis.
15–35%
range of over-sizing found, because most pools were sized to headcount, not peak logins.
1 pool
shared by every MAS application, so unused points can move to the apps that need them.
Sizing figures from Redress Compliance, IBM Maximo Application Suite licensing analysis.
Users draw AppPoints when they log in, at a rate set by their entitlement level, and return them when the session ends. Each user is charged at the highest level their security groups allow, so the cost of adding an application depends on who gets access to it.
| Application | User entitlement needed | AppPoints per concurrent user |
|---|---|---|
| Maximo Monitor | Limited | 5 |
| Maximo Health | Base | 10 |
| Maximo Predict | Premium | 15 |
Entitlement levels per IBM documentation and third-party licensing guidance. IBM revises tier contents between releases; confirm against your own order documents.
Estimate your pilot’s peak draw
Enter how many people would be logged in to each application at the same time.
A worked example
A pilot team of three planners on Health and two reliability engineers on Predict, all logged in at once, draws 3 × 10 + 2 × 15 = 60 AppPoints at peak. Users who already hold Base or Premium entitlement for other work add nothing. For many clients, that fits inside the headroom they already have.
05The pipeline
Sensor to Maximo in five stages
DIME sits between the equipment and Maximo. It runs at the edge, close to the machines, and turns raw signals into asset-aware data that each Maximo application can use.
-
01
Connect
What DIME does
Reads directly from PLCs, CNCs, robots, and meters through native connectors, including OPC UA and DA, Modbus TCP, Ethernet/IP, Siemens S7, Beckhoff ADS, MTConnect, Fanuc, Haas, SNMP, MQTT, and SparkplugB.
What Maximo gets
Data from mixed-vendor and legacy equipment, without a driver project per machine
-
02
Contextualize
What DIME does
Maps each tag to its Maximo asset, site, and meter, converts units, and combines signals with Lua or Python scripts.
What Maximo gets
Readings that land on the right asset record, in the units Maximo expects
-
03
Reduce
What DIME does
Reports by exception, sending values only when they change, and rolls high-rate signals up into meaningful intervals.
What Maximo gets
Clean meter history, without flooding Manage or the IoT layer
-
04
Detect
What DIME does
Learns each asset’s normal pattern continuously with Hierarchical Temporal Memory (HTM) and scores how unusual current behavior is.
What Maximo gets
An anomaly score per asset that can be trended, alerted on, and reviewed
-
05
Deliver
What DIME does
Publishes readings and anomaly scores to Monitor over MQTT and to Manage through its REST APIs, and feeds Splunk and YouBIM from the same stream.
What Maximo gets
One consistent feed to every system that needs it, from a single configuration
One configuration, under version control.
Each DIME pipeline is defined in a readable YAML file. Teams can generate a first draft from a plain-English description with AI, validate it against the published schema, review it like code, and roll it out to the next line or plant.
Native Maximo integration in the Scale edition
The DIME Scale edition includes native integration with Maximo, Splunk, and YouBIM, along with Kepware and legacy migration support. DIME can run alongside existing OPC servers and historians during a transition, so plants do not have to switch everything at once.
06Closing the loop
From anomaly to work order, and back again
Data in Maximo is only half the loop. The return comes when a deviation becomes planned work, and the outcome of that work makes the next detection better.
-
Stage 1
Detect
DIME flags a sustained deviation from the asset’s learned normal, at the edge, in real time.
Maximo record
Meter reading or anomaly measurement
-
Stage 2
Confirm
Monitor, or a condition-monitoring limit in Manage, evaluates the signal against rules set by reliability engineers.
Maximo record
Alert or service request
-
Stage 3
Act
A planner reviews the alert and creates a work order from a standard job plan, with history and parts in view.
Maximo record
Work order with job plan
-
Stage 4
Learn
The technician closes the work with failure codes and actuals. Health, Predict, and future reviews draw on that record.
Maximo record
Failure history and actuals
Guardrails against alarm fatigue
- Alert on sustained anomalies, not single readings.
- Route alerts to review before they become work orders.
- Start with critical assets and known failure modes.
- Track the share of alerts that led to real findings.
What planners see in Maximo
- The asset, its current readings, and its trend.
- Open and recent work orders on the same asset.
- The anomaly score that raised the alert.
- In YouBIM, where the asset sits and how to reach it.
07Where HTM fits
Fewer, better alerts with continuous learning
Most condition alerts come from fixed thresholds or from models trained on a snapshot of history. Both have a place. Both struggle when normal operation keeps changing.
| Approach | How it decides | What it asks of you |
|---|---|---|
| Static thresholds | Fires when a value crosses a fixed limit. | Living with noise on startups and changeovers, and keeping limits current by hand |
| Periodically trained models | Learns from a historical training set, then scores new data against it. | Retraining as equipment, products, and seasons change, and the data science time to do it |
| Continuous learning (HTM in DIME) | Learns sequences of normal behavior for each asset, updates with every reading, and scores how far current behavior departs from what it expected. | A learning period on each asset, and a decision on how long a deviation must persist before it alerts |
HTM does not replace Maximo Predict. Predict forecasts failures from history; DIME shows that behavior has changed right now. Used together, the DIME anomaly score becomes one more signal for Monitor and Predict, and the alerts that reach planners are ones worth investigating.
08Architecture
Built for OT security and IT governance
The pipeline has to pass review by the controls engineers who own the equipment and the IT team that owns the network.
09The path forward
Start with one asset class and one failure mode
The fastest route to value is narrow and deep: a small group of critical assets, a failure mode the team already understands, and a complete loop from sensor to closed work order.
Phase 1
Pilot: first 90 days
- Run an AppPoints Utilization Review to size the pilot team.
- Pick 5 to 10 critical assets with a known, costly failure mode.
- Connect them with DIME and map tags to Maximo assets and meters.
- Let HTM learn normal; set review rules in Monitor or Manage.
- Measure alerts raised, findings confirmed, and work orders created.
Phase 2
Scale: 6 to 12 months
- Turn the tag-to-asset mapping into a reusable template.
- Extend to the next asset class, then the next plant.
- Feed Health scoring and Predict models with live history.
- Manage every site from one control plane with Zenith.
- Rebalance AppPoints at renewal based on measured use.
Measure what matters
Alert precision
The share of alerts that led to a confirmed finding.
Detection to work order
The time from first anomaly to a planned work order.
Planned vs. reactive work
The trend on the assets in scope, month by month.
Unplanned downtime and MTBF
Measured against your own baseline, not an industry average.
10Readiness scorecard
Is your Maximo ready for live machine data?
Score one point for each statement that is true today. Your score shows where to start.
0/ 10
Your score
-
8–10
Ready to connect
Pilot the full loop on critical assets now, and plan the second plant.
-
4–7
Connect and fix together
Start connecting while you close gaps in meters, failure codes, and job plans.
-
0–3
Foundation first
Strengthen the Manage foundations first. A Data Health Check is the right next step.
11Why SMS
One partner from the machine to the work order
Strategic Maintenance Solutions has helped organizations get more from IBM Maximo since 1999. With DIME, we deliver the data layer and the system of record as one program.
25+
Years of expertise
1,000+
Successful projects
F1000
Client base
2
Continents
- IBM Platinum Partner. Deep experience across Maximo Manage and the full Maximo Application Suite, including Monitor, Health, and Predict.
- DIME, delivered by SMS. DIME is built by MRIIOT and delivered by SMS. The Scale edition includes native Maximo, Splunk, and YouBIM integration, Kepware and legacy migration, and air-gap and high-availability support.
- From sensors to field services. Predictive maintenance field services, condition monitoring, and panel builds from the same team that configures Maximo.
Your next step
Request an AppPoints Utilization Review
A free review that shows what you own, what you use, and where to start:
- Your AppPoint entitlement compared with measured use
- The AppPoints Monitor, Health, and Predict would draw for your team
- Critical assets, controllers, and Maximo meters to connect first
- A DIME pilot scope to feed those assets’ live data into Maximo
- A rollout path from the pilot to the next asset class and plant