New England Database Society Friday, September 24 sponsored by Sun Microsystems
NEDS

The Design of Stream Mill: A Data Stream Management System of Many Uses

Carlo Zaniolo  
UCLA

Friday, September 24, 2004, 4:00 PM
Volen 101, Brandeis University

(preceded by a wine and cheese reception at 3:00 pm)

Abstract:

This talk presents a database-inspired approach to the design of efficient data stream management systems for a broader application spectrum: e.g., we want to support data stream mining, sequence queries, and streaming XML queries. Toward this goal, we must overcome the problems that (i) currently, SQL does not support these applications well, even on database tables, and (ii) its power is further reduced by the constraint that blocking query operators must be avoided on data streams. We present an in-depth analysis of these problems, and discuss how our Expressive Stream Language (ESL) solves them. The main ESL extension w.r.t. SQL is that new aggregates can be natively written in SQL, which thus become Turing-complete. Furthermore, both the base version and the window version of each aggregate can be optimized to support advanced applications efficiently. The Stream Mill server supports continuous queries and ad-hoc queries using a reconfigurable execution model and highly adaptive scheduling strategies.

Speaker Bio:

Carlo Zaniolo grew up in Italy, and graduated from Padua University in Electrical Engineering. He got his PhD in CS from UCLA in 1976 with the thesis: "Analysis and Design of Relational Schemata for Database Systems." After that, he worked for industrial research labs in New England (Sperry Research), New Jersey (Bell Labs), and Austin, Tx (MCC). Then, in 1991 he returned to UCLA where he occupies the N.E. Friedmann Chair in Knowledge Science. He served as the program (co)chair of EDBT, SIGMOD and VLDB--in the fabled old times when most of the good papers were accepted.

Maintained by Dina Goldin dqg AT cse.uconn.edu
Last updated on 04/13/04