Publication Details
An Executable Sequential Specification for Spark Aggregation
Chen Yu-Fang (ASIN)
Lengál Ondřej, Ing., Ph.D. (DITS FIT BUT)
Mu Shin-Cheng (ASIN)
Sinha Nishant (IBM)
Wang Bow-Yaw (ASIN)
Data Parallel Computation, Functional Specification, Requirements, Verification, Spark
Spark is a new promising platform for scalable data-parallel computation. It provides several high-level application programming interfaces (APIs) to perform parallel data aggregation. Since execution of parallel aggregation in Spark is inherently non-deterministic, a natural requirement for Spark programs is to give the same result for any execution on the same data set. We present PureSpark, an executable formal Haskell specification for Spark aggregate combinators. Our specification allows us to deduce the precise condition for deterministic outcomes from Spark aggregation. We report case studies analyzing deterministic outcomes and correctness of Spark programs.
@INPROCEEDINGS{FITPUB11330, author = "Chih-Duo Hong and Yu-Fang Chen and Ond\v{r}ej Leng\'{a}l and Shin-Cheng Mu and Nishant Sinha and Bow-Yaw Wang", title = "An Executable Sequential Specification for Spark Aggregation", pages = "421--438", booktitle = "Proceedings of NETYS'17", journal = "Lecture Notes in Computer Science", number = 10299, year = 2017, location = "Heidelberg, DE", publisher = "Springer Verlag", ISSN = "0302-9743", doi = "10.1007/978-3-319-59647-1\_31", language = "english", url = "https://www.fit.vut.cz/research/publication/11330" }