Collaborative Stabilization

Abstract

In this paper, we present the paradigm of collaborative stabilization that focuses on providing stabilization in the presence of an essential but potentially disruptive environment. By essential, we mean that without the environment actions, stabilization property would be impossible. At the same time, environment actions are not in the control of the program and can be disruptive to the recovery. We demonstrate the need for collaborative stabilization by providing examples where existing paradigms of stabilization are undesirable/insufficient. We compare collaborative stabilization with existing paradigms of stabilization. We identify the complexity of verifying collaborative stabilizing programs and develop theorems that focus on composition of such programs.