Ideal Stabilization in a PIF Chain

I test the practicality of ideal stabilization. Self-stabilization, or classic stabilization, is a specification property which guarantees that a program will eventually satisfy its specification regardless of transient faults or lack of initialization. Classic stabilization loses some appeal as a new standard for fault tolerant distributed systems due to unattractive features such as poor composition and uncontrolled behavior during fault recovery. Ideal stabilization is a special case of stabilization which eliminates these two disadvantages. However, ideal stabilization has mostly been discussed in theory and needs practical testing. I test this property in a real program with real faults to observe the finer details of its behavior in a practical setting. Specifically, I create a simulation engine to test the behavior of a program with the propagation of information with feedback (PIF) specification. This specification is proven to ideally stabilize, and is a well-known model for message passing and clock synchronization in distributed systems. The PIF specification usually comes in the form of a tree, but for simplicity I reduce this tree to a chain. I present this program with initial faults as well as totally random initial states, and observe its behavior in regards to fault tolerance during computational operation.