CrocodileDB: Efficient Database Execution through Intelligent Deferment

The end of Moore’s law will push database system designers to be more judicious with computation as the growth in data outpaces the availability of computational resources. Eagerness, or aggressively consuming resources to immediately and quickly complete the task at hand, is one source of wasted resources in modern data systems where the systems expend unnecessary resources waiting on queries, data, or both. Intelligently deferring a task to a later point in time can increase result reuse, reduce work that might later be invalidated, or avoid unnecessary work altogether. We propose a research prototype system, CrocodileDB, which is a resource-efficient database system that automatically optimizes deferment based on user-specification and workload prediction. CrocodileDB integrates new ways of specifying timing information, new query execution policies, new task schedulers, and new data loading schemes.