FlumeJava: Easy, Efficient Data-Parallel Pipelines

Abstract
1. Introduction
2. Background on MapReduce
3. The FlumeJava Library
  3.1 Core Abstractions
  3.2 Derived Operations
  3.3 Deferred Evaluation
  3.4 PObjects
4. Optimizer
  4.1 ParallelDo Fusion
  4.2 The MapShuffleCombineReduce (MSCR) Operation
  4.3 MSCR Fusion
  4.4 Overall Optimizer Strategy
  4.5 Example: SiteData
  4.6 Optimizer Limitations and Future Work
5. Executor
6. Evaluation
  6.1 User Adoption and Experience
  6.2 Optimizer Effectiveness
  6.3 Execution Performance
7. Related Work
8. Conclusion