Object Orientation Meets Big Data - Performance Impact, Restoration, and Thoughts on Language Design

  主讲人: Prof. Harry Xu, Associate Professor, University of California, Irvine




  Object orientation (OO) is a powerful programming methodology that lays the foundation for most of today's large-scale software applications, including data-intensive systems such as Spark and Hadoop, which have increasingly large impact on our daily lives. OO contains a set of abstractions that provide the benefit of modularity and simplicity in programming at the cost of a performance penalty at run time. While this penalty is arguably small for non-data-intensive applications, it can get significantly magnified in ''big data'' systems that create and manipulate billions of data objects, becoming too large to accept in production settings.


  Harry Xu is an Associate Professor in the Computer Science Department of University of California, Irvine. Harry worked at Microsoft Research as a Visiting Researcher in 2017, where he created and led the development of a project that aims to build an optimizing compiler for multilingual data analytical pipelines, and in particular, Microsoft’s Scope/Cosmos. He worked at IBM T. J. Waston Research Center as a Co-op/intern from 2008 to 2011 where he led the development of a series of runtime bloat detection tools. His research ranges from software engineering, through programming languages and compilers, to runtime/operating/distributed systems and computer architecture. Harry is the recipient of the AITO Dahl-Nygaard Prize (2018), a Distinguished Research Award from Ohio State University (2011), as well as a number of Distinguished Paper Award from ACM SIGPLAN and SIGSOFT.