CS240B
Spring 2008
CS240B
ADVANCED DATA BASES and KNOWLEDGE BASES
Instructor: Carlo Zaniolo
Office Hours: Tuesdays 3:00--5:00pm
Research Presentations
June 4
Burst Detection in Data Streams
Association Rule Mining by Cheuk Yu Yeung
May 28
Gigascope: A Stream Database for Network Applications by Swathi Ganapathi
Minining for Frequent Sequences & Closed Itemsets by Grace Shih
May 21
XML Query Systems by Yung-Cheng Chen
Load Shedding Techniques in Aurora by Zhen Huang
May 19
Continuous Query Support (in DBMS and DSMS) by Alexander Shkapsky
Clustering Data Streams: Birch, Stream, Clu-Stream by MyungWon Ham
May 14
Filtering XML Documents with XPath by Nickolaus Phan
XPath on Streaming XML using YFilter and Xaos by Daniel Carney
May 12
Density-Based
Clustering & Visualization of Cluster Structure
By Nikolay Pavlovich Laptev
Week 4. Assignment (Due on WD, April 30)
Download the Weka system and use it to do the following project.
Week 3. Assignment (Due Monday, April 21)
Although DSMS are now taking a leadership role in the Complex Event Processing (CEP) fiel, many other approaches are actively being pursued by commercial vendors and researchers. Your assignement is to compare the following CEP systems and products
- DSMS startups, including, Streambase, Coral8, Apama, and Truviso (and any other startup company you might know or discover through the web, e.g., Oracle CEP extensions)
- Systems and product that support CEP using languages other than SQL and XML. For that you might find the following references useful:
- www.unix.com/blogimgs/EVENT.PROCESSING.LANGUAGE.SURVEY.V14.OCT.15.2006.pdf
- Dagstuhl Seminar Proceedings 07191: Event Processing
- The Power of Events: An Introduction to Complex Event Processing in Distributed Enterprise Systems by David Luckham
Please write a short report (1400 words or less) to describe the history, techniques and application forte of the different approaches to CEP, and the systems implementing them.
Week 2. Assignment. Solutions can be found in here.
Read The following three papers:
- A Data Stream Language and System Designed for Power and Extensibility. Yijian Bai et al., ACM 15th Conference on Information and Knowledge Management (CIKM'06), Arlington, Virginia, Nov 6-11, 2006.
- A heartbeat mechanism and its application in gigascope. by Theodore Johnson et al.: In VLDB 2005, pages 1079-1088.
- Optimizing
Timestamp Management in Data Stream Management Systems. by Yijian
Bai et al.: In ICDE 2007.
Task 3.2. We showed how to fix the idle-wating problem for union and joins. However, aggregates on logical windows with slides can also experience a similar problem- explain how the problem can occur (you might want to start from the case of tumbles)
- give the execution of aggregates on a logical window with slides
in the style used for union and joins in the above paper Yijian
Bai et al.:
Week 1. Assignments. Solutions can be found in A, B, and C.
* Study K. Guion's overview on OLAP Functions.
* Study the proposed Match-Recognize ANSI Standards for SQL (SQL-MR for short) Discussion blog on SQL-MR
* Study Models and Issues in Data Stream Systems. by B. Babcock, S. Babu, M. Datar, R. Motwani, and J. Widom. Invited paper in Proc. of PODS 2002, June 2002.
Task 1.1. Given the temporal table:
emp-history( E#, Dept, Sal, Tstart, Tend)
Use SQL:2003 OLAP functions to compute the relation
emp-history( E#, Dept, Tstart, Tend)
obtained from the previous relation by projecting out Sal and coalesching the time intervals ( You can find a solution of the similar problem in these slides and this paper)
Task 1.2. Express in SQL-MR the example query the SQL-TS queries in the slides
Task 1.3. Write a short (one page or less) report on SQL-MR explaining the parts that, in your view, have problems or require clarification (e.g., my view is thatoutputs are not clear in the presence of alternative patterns).