Park 'N Park
Park 'N Park is a fault-tolerant, distributed, real-time three-tiered application for tracking parking garage usage. This was an academic project for the "Fault Tolerant, Distributed, Real-Time Systems" course taught by Associate Professor Priya Narasimhan during the Spring semester of 2006. Park 'N Park manages parking lots by keeping track of how many spaces are available in lots and recommends other lots when the lot a driver is at is full.
This project was the capstone of the course and required the team to implement a multi-tiered, distributed system that can tolerate faults within a deterministic time to achieve real-time guarantees within specific constraints. To achieve this, the focus of the project and most of the software code implements fault handling and recovery for all points of failure in the system at and above the business logic server tier. Database disconnections and crashes are gracefully handled, though database redundancy and replication were not implemented in the project. As a team, we measured response times within the application as well as response times during faults and fault recoveries to ensure that we were meeting our real-time performance goals.
More information about this project exists on its web site at Carnegie Mellon University, including roles, database schema, and the source code.