Gumbo: Guarded Fragment Queries over Big Data

We present Gumbo, a system for the ecient evaluation of guarded fragment queries on top of Hadoop and Spark. A key asset of Gumbo is the reduced number of jobs in comparison with recent systems such as Pig, Hive or Shark. For unnested guarded fragment queries, Gumbo even provides a constant bound on the number of jobs independent of the size of the query. In the demo, we will address the following features of Gumbo: ease-of-use, query plan construction and visualisation, and query execution.