M1. | Install Hadoop on a SnowFlock-based virtualized environment. |
M2. | Implement a Java wrapper for the SnowFlock Python or C APIs. |
M3. | Identify modification required to extend the Apache Hadoop code with the ability to dynamically allocate virtual machines. |
M4. | Implement the identified extensions. |
M5. | Identify the applications to be used for performance measurements. We aim at benchmarking common MapReduce use-cases, such as tuple selection in the map step and aggregation in the reduce step. |
M6. | Execute identified tests and analyze the results. |