Scylla User Guide

To translate a BPMN diagram into the desired XES log, one can make use of the Scylla process simulator. In order to be able to use our implementation one has to pay attention to a couple of specialties:

• The simulation is executed on the blockchain in ❗real time❗, not in the short time that Scylla needs for simulating the processes. Keep this in mind when running a scenario on the blockchain. Recommended are scenarios that take about half an hour.
• The data to be stored on the blockchain has to be modelled as BPMN data objects. This also comes with a small adjustment in the simulation configuration file, mentioned below.

For each writing activity to a data object one has to specify what is written to it. In our approach the what isn't as relevant, as we store only mock data on the chains. More interesting for our use case is the payload of the data object. For our approach to work, one should therefore specify it as follows in the simulation configuration file:

<bsim:dataObject id="DataObjectReference_id" name="DataObject Name">
      <bsim:field name="payload" type="double"> <!-- Define a field with the name payload and a type of double-->
        <bsim:arbitraryFiniteProbabilityDistribution>
          <bsim:entry value="30" frequency="1"></bsim:entry>
        </bsim:arbitraryFiniteProbabilityDistribution>
      </bsim:field>
</bsim:dataObject>

The distributions of what is written in the payload field can and should be adjusted, please note, that payload sizes bigger than 200 bytes can not be handled by the backend. Our benchmarking systems works as long as each initialised data object has a payload field. The payload should be defined in bytes.