Abstract
We design and implement a linear hash algorithm in nested transaction environment to handle large amount of data with increased concurrency. Nested transactions allow parallel execution of transactions, and handle transaction aborts, thus provides more concurrency and efficient recovery. We use object-oriented methodology in the implementation which helped in designing the programming components independently. In our model, buckets are modeled as objects and linear hash operations are modeled as methods. The papers contribution is novel in the sense that the system, to our knowledge, is the first to implement linear hashing in a nested transactions environment. We have build a system simulator to analyze the performance. A subtle benefit of the simulator is that it works as the real system with only minor changes.
