Kagera. A Discrete, colored Petri Net DSL & Executor.

Example of a simple colored petrinet with functions acting on the data in the tokens.

val sum = {

    // places
    val a = Place[Int](id = 1, label = "a")
    val b = Place[Int](id = 2, label = "b")
    val c = Place[Int](id = 3, label = "result")

    // transitions
    val init = Transition(id = 1, label = "init") {
      () ⇒ (5, 5)
    }

    val sum = Transition(id = 2, label = "sum") {
      (a: Int, b: Int) ⇒ a + b
    }

    // process topology
    process(
      init    ~> %(a, b),
      %(a, b) ~> sum,
      sum     ~> c)
  }

Transition Model

Input:

• M is the in-adjacent marking of the transition.
• S is the type of state the transition closes over, Unit indicates no state.
• I Is the type of input the transition requires (provided from outside the proces), Unit indicates case no input is required.

Output:

• M' is the out-adjacent marking of the transition.
• E Is the type of event or output the transition emits. Unit indicates no event / output.

The transition's behaviour is defined by 2 functions

1. (M, I, S) => (M', E) A function producing the out-adjacent marking and event from in-adjancent marking, state and input.
2. S => E => S An event sourcing function which updates the state using the emitted event / output object.

Roadmap

Short term:

• Http API (using akka http)
• Clustering & Sharding
• Analysis (using apache Spark or apache Flink), some examples:
  • Path querying
  • Path prediction
  • Transition timing

Long term:

• Hyarchical petri nets
• Timed petri nets
• Other execution models, such as:
  • Splitting processes over multiple actors, possibly on different nodes in a cluster.
  • A petri net process as a Stream (using https://github.com/functional-streams-for-scala/fs2 or akka streams for example).
• Process migration
  • Providing tools to migrate process state when the petri net topology changes.
• Visual process editor