Long-running transformations

When a transformation involves multiple steps, with the periodic addition of input products and periodic extraction of output products, the transformation can be recorded as a series of transformation events. These transformation events are linked with a Transformation ID that identifies which transformation events are part of the same long-running transformation. Long-running transformations are highly relevant for process manufacturing, for example, the beverage, paint, and chemical industries.

Modeling long-running transformations

Long-running transformations can be modeled in two different ways:
  • Model 1 - As a single EPCIS Transformation event
    A single EPCIS Transformation event that lists all input products and all output products from the entire time period.
  • Model 2 - As multiple EPCIS Transformation events
    Multiple EPCIS Transformation events that maintain the relationship between the input products and output products from all events that are linked with the Transformation ID.
For more information about EPCIS and long-running transformations see, EPC Information Services (EPCIS) Standard and EPCIS and CBV Implementation Guideline.

Example long-running transformation

For example, consider the supply chain process of creating wine, where several transformations, like crushing grapes, fermenting, filtering, and aging, are necessary to produce the final product. In this example, as the wine ages, wine is extracted different time points, recorded with transformation events. These transformation events are part of the same long-running transformation, which are linked with the Transformation ID Wine123.

Figure 1. Example of wine creation, including a long-running transformation
To produce wine, first grapes are crushed, then the crushed grapes are fermented, and then the fermented grapes are filtered. Next, the wine is aged and extracted at two different time points. The transformation events that record the extraction of aged wine are part of the same long-running transformation and are linked with the Transformation ID, Wine123. Both events include Filtered wine 900L as the input product. Event A4 includes Beaujolais Nouveau 200L as an output at time 08/02/2023, 9am. Event A5 includes Cabernet Sauvignon, 300L as an output at time 01/05/2023, 9pm.

Model 1 - Modeling as a single EPCIS Transformation event

Long-running transformations can be modeled as a single EPCIS Transformation event. This transformation event lists all input and all output EPCs that were involved in the transformation and typically uses the time that the process completed as the event time. Modeling long-running transformations as a single transformation event can work if the output EPCs are produced at the same time.

However, this model doesn’t differentiate input EPCs or output EPCs that are added or removed at different points in time, since a single timestamp is recorded for the entire transformation. Similarly, with a single event, all outputs are associated with the same instance/lot master data (ILMD), so differences in the output lots, like expiration dates cannot be differentiated. In addition, restriction on the number of EPCs that can be included in a single transformation event can also limit the use of this model.

For example, if the aging wine (Figure 1) is modeled as a single transformation event, information about the separate batches of wine is lost. Specifically, the event time that Beaujolais Nouveau, 200 L was extracted (Table 1).
Table 1. Long-running transformation of aging wine, modeled as a single EPCIS Transformation event
Description Aging wine
Event type Transformation
Input EPCs Filtered wine, 900 L
Output EPCs Beaujolais Nouveau, 200 L
Cabernet Sauvignon, 300 L
Event time 01/05/2023, 9 PM

Model 2 - Modeling as multiple EPCIS Transformation events

Long-running transformations can be modeled with multiple EPCIS Transformation events. Modeling long-running transformations with multiple EPCIS Transformation events is preferred if output EPCs are extracted at different time points or input EPCs are added at different time points. Specifically, the relationship between input EPCs or output EPCs which are added or removed at different points in time is preserved with the use of multiple transformation events.

For example, if the aging wine (Figure 1) is modeled as multiple transformation events, information about the separate batches of wine is maintained (Table 2).

Table 2. Long-running transformation of aging wine, modeled as a series of EPCIS Transformation events
Description Aging wine (Event ID: A4) Aging wine (Event ID: A5)
Event type Transformation Transformation
Input EPCs Filtered wine, 900 L Filtered wine 900 L
Output EPCs Beaujolais Nouveau, 200 L Cabernet Sauvignon, 300 L
Event time 08/02/2023, 9 AM 01/05/2023, 9 PM

Comparison of Model 1 and Model 2

In the wine example, using multiple EPCIS Transformation events is preferred, since the two batches of wine are produced at different time intervals. Using a single transformation event doesn’t maintain the relationship between inputs and outputs and records a single timestamp for the entire transformation. Table 3 provides a comparison between modeling with a single EPCIS Transformation event and modeling with multiple EPCIS Transformation events.

Table 3. Comparison of modeling long-running transformations with a single EPCIS Transformation event vs multiple EPCIS Transformation events
Model 1 - A single EPCIS Transformation event Model 2 - Multiple EPCIS Transformation Events
  • All inputs and outputs are associated to a single timestamp.
  • All outputs are associated with the same ILMD.
  • The number of EPCs that can be included in a single transformation event is restricted to 3500.
  • Works when the outputs are produced at the same time and are not subject to further business steps.
  • The relationships between inputs and outputs are maintained.
  • Output lots can be associated to different ILMDs.
  • Preferred model when outputs are generated over disparate time intervals and outputs are subject to further business steps before the transformation completes.

For more information about EPCIS and long-running transformations see, EPC Information Services (EPCIS) Standard and EPCIS and CBV Implementation Guideline.

Long-running transformations in Transparent Supply

With Transparent Supply, you can model long-running transformations with a single EPCIS Transformation event (Model 1) or multiple EPCIS Transformation events (Model 2). By default, the Trace user interface does not display long-running transformations with multiple EPCIS Transformation events that are linked with a Transformation ID (Model 2), but provisioning is available to enable Model 2. To discuss enabling long-running transformations (Model 2) for your organization, open a case with IBM Support.