Blog / MES
MES · Guide

MES vs DCS: Control Loop and Batch Record

OL By OEE Lab Editorial |Updated 2 November 2026 |6 min read
Short answer

A distributed control system (DCS) controls a process: it runs the loops and sequences that hold temperature, pressure and flow, often many times a second. An MES manages production around it: orders, materials, batch records, quality and performance. The DCS keeps the process in spec, and the MES shows why an order ran late.

This comparison comes up in chemical, food, pharma, refining and utility plants. In those places, the DCS is often already the biggest system on site, and someone asks whether it can do the MES job too.

The answer depends on one distinction: controlling a process and managing production are separate problems, with separate failure costs.

What a DCS is for

A DCS spreads controllers across a plant and ties them to operator stations and an engineering environment, all in one vendor-supplied system. Its controllers run PID loops, interlocks and sequences at Level 2 of ISA-95, just above the sensors and valves at Level 1.

It is built to keep running. Redundant controllers, networks and power are common, because losing control of a reactor is a safety event.

Control timing is measured in fractions of a second to seconds. Change management is strict, since a bad edit can hurt people or equipment.

What an MES is for

An MES works at Level 3, in minutes to shifts. It decides or receives which order runs in which unit, checks the materials and the people, collects the record and reports the results.

If the MES goes down, the process must carry on safely under the DCS. That one rule shapes the whole design: the MES can ask, and the DCS decides.

Task in a batch plantUsually owned by
Hold reactor at 82 C with a PID loopDCS
Run phases in order: charge, heat, react, coolDCS or its batch manager
Interlock: no heating without agitationDCS, with a safety system for safety functions
Which order goes in which reactor, and whenMES, from the plan
Check raw material lots against the recipeMES
Set recipe quantities for this batch sizeMES sends them, DCS applies limits
Batch record, deviations, genealogyMES, using data from the DCS
Losses and OEE per unit and shiftMES
A common split. Exact lines vary by plant and sector, so agree on them before design starts.

ISA-88 is where they connect

ISA-88, the batch control standard, gives both sides a shared language. It separates a recipe from the equipment that runs it, and it breaks the procedure down into unit procedures, operations and phases.

In one common setup, the MES holds the order and the formula values, then creates the batch. The DCS or its batch manager runs the phases that contain the actual control logic.

A plant with a strong DCS batch engine may keep more of the sequence there, while pharma plants often run more of the manual steps in MES electronic batch records. See MES for pharma.

The data that passes between them

  • MES to DCS: batch ID, recipe or formula values, batch size, start permission.
  • DCS to MES: phase start and end times, actual quantities, final parameters, alarms and deviations.
  • In and out of a historian: the DCS sends process tags to a historian. The MES reads curves and writes batch context, as MES vs historian describes.

The links usually run over OPC UA or a vendor interface, through a DMZ or a firewalled zone. Most plants keep MES writes to a short, reviewed list of tags.

Example: the batch the DCS thinks went well

Example: a batch reactor has a planned cycle of 360 minutes. The batch runs every phase exactly on time, and the DCS report looks clean.

Then the finished product waits 45 minutes in the vessel for a quality sample result, so the unit is tied up for 405 minutes.

Charge30 min
Heat60 min
React180 min
Cool60 min
Wait for QA result45 min
Discharge30 min
Example batch, 405 minutes in total. The DCS sees five phases. The red bar is the wait that only a production record holds.

The 45 minutes never appear as a fault on the control system, because nothing went wrong with the control. On an order level they are a real loss of capacity.

Availability for that batch window would be 360 / 405 = 88.9%, if you treat the wait as unplanned. How to count these holds varies by plant, as OEE for batch processes explains.

The lab side of that wait is covered in MES vs LIMS. You can also try your own numbers in the availability calculator.

Can a DCS do the MES job?

Part of it. Modern systems offer batch modules, reports and production dashboards, and those can satisfy a plant with simple needs.

Check these before you decide:

  • Manual work. Weighing, additions and sign-offs need user prompts, lot scans and audit trails that may sit outside a control system.
  • Cross-unit view. One DCS often covers one area, while an order may cross packaging lines on PLCs.
  • ERP link. Control engineers usually do not want a business interface in the control domain.
  • Change control. Every report change in the DCS goes through the strict control change path. That is slow for production reporting, and it should be.

DCS and MES in a continuous plant

A continuous unit has no batches, so the MES role shrinks. Plants there often use production accounting, a historian and lab data, sometimes grouped as manufacturing operations management.

The useful questions become grade changes, off-spec time, and rate against the design rate. See OEE for continuous processes and OEE for process manufacturing for how the factors apply.

Which setup fits which plant

PlantUsual setupWhy
Continuous, one or few productsDCS and historianStable operation, with little tied to single orders
Batch, multi-product or regulatedDCS and MESEach lot needs a record, and each order needs a slot
Discrete or packaging linesPLCs and MESNo process loops to run, and the question is lost time
Simple manual batch workPaper batch sheetFew signals, and few decisions to support
Typical patterns, not rules. Packaging after a process step often pulls an MES into a DCS plant.

For the control side of a discrete plant, see MES PLC integration. A paper batch sheet and a simple tracker can still give you useful numbers.

Four rules for the boundary

  • Keep business logic out of the controllers. Scheduling rules inside a DCS become very expensive to change.
  • Keep control logic out of the MES. If a sequence fails when the MES server restarts, the design is wrong.
  • Limit writes from Level 3. Each writable tag needs a reason, a range check on the DCS side and a log.
  • Keep one recipe master. A value edited in the DCS and not in the MES leaves the batch record wrong.
See it live in Fabrico

A control system keeps the process right, but it will not tell you where the capacity went on packaging or discrete lines around it. Fabrico is a light MES for OEE: stops, speed and counts come live from PLCs or IoT sensors, with OEE per line and shift. See how Fabrico's light MES for OEE works.

What people ask next

Is a DCS the same as a PLC?

They have converged, but a PLC is traditionally a controller for machines and a DCS is a whole plant control system with built-in operator stations and engineering. Both sit below the MES in ISA-95.

Is a DCS the same as SCADA?

Not quite. A DCS closes control loops inside its own integrated system and suits process plants. SCADA supervises PLCs or remote sites and focuses on monitoring and data, with control done in those devices.

Does an MES ever control the process directly?

It should not. An MES can send a recipe, a setpoint or a start permission, and the DCS runs and limits the control. A failed MES should have no effect on process safety.

Can you calculate OEE on a DCS plant?

Yes. Run state, rate and off-spec time are usually available as tags, and planned time comes from production records. See OEE for continuous processes for the factor definitions.

What is an electronic batch record?

It is the digital record of how a batch was made, including materials, steps, parameters, sign-offs and deviations. It usually lives in the MES and draws process values from the DCS. See MES for pharma.

See OEE per line and shift on your own machines

Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.

Book a Fabrico demo

← More OEE Lab guides · OEE for batch processes →

OEE essentials: OEE calculator · What is OEE · Six big losses · Availability · TEEP · Cost of downtime · OEE benchmark · Free OEE tracker · Free OEE software · How to track downtime · How to implement OEE · Light MES vs full MES