Monte Carlo Methods for Managing Uncertain Enterprise Data

Fei Xu
                                                                  Microsoft
                                                                  Research

Date: May 24, 2010 (Monday)
Place: E & C S BUILDING Second Floor Room 2120
TIME: 10:00 (DONUTS) 10:10 (TALK)

Modern enterprises must manage uncertain data for purposes of risk assessment
and decision making under uncertainty.  In this talk, I will first discuss
MCDB (the Monte Carlo Database System), a system for managing uncertain data
that is based on a Monte Carlo approach.

The Monte Carlo approach is well suited for managing uncertain enterprise
data. MCDB can support industrial strength business-intelligence queries over
uncertain warehouse data. Moreover, MCDB's extensible approach to specifying
uncertainty can also capture complex stochastic prediction models, allowing
sophisticated ``what-if'' analyses within the DBMS. However, the Monte Carlo
computations can be highly CPU intensive. But it offers the potential for
massive parallelization. To realize this potential, we developed a new
system, called MC3 (Monte Carlo Computation on a Cluster), that extends
the MCDB approach to the map-reduce processing framework. MC3 can exploit
the robustness and scalability of map-reduce, and can handle data stored
in non-relational formats. In the second half of the talk, I will discuss
the new system. (Joint work with Peter Haas, Kevin Beyer,  Vuk Ercegovac,
Bo Shekita, Chris Jermaine, Ravi Jampani, Luis Perez, Mingxi Wu)



Biography:
Fei Xu obtained his PhD degrees from University of Florida in 2009. Since then
he has been a research software development engineer in Bing Infrastructure
Team Microsoft. His research interests are in large scale data management. He
is especially interested in managing data uncertainty, Cloud Computing and
managing biological data