Accountability in computing implies that an entity should be held responsible for its behaviors with verifiable evidence. In order to study accountability, quantitative methods would be very helpful. Even though there are some researches in accountability, there are no other works which study quantitative accountability in practical settings, while quantitative accountability is defined as using quantities or metrics to measure accountability. In this paper, we propose P-Accountability, which is a quantitative approach to assess the degree of accountability for practical systems. P-Accountability is defined with two versions, a flat model and a hierarchical one, which can be chosen to use depending on how complex the system is. We then provide a complete case study that applies P-Accountability to PeerReview, which provides Byzantine fault detection for distributed systems. In addition, we propose Traceable PeerReview, which is our effort to apply PeerReview to wireless multi-hop environments. In addition, through the system evaluation we can show that the simulation outcomes are aligned with the numeric results.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Electrical and Electronic Engineering