CSED700 Advanced Topics in Data Management

1
CSED700 Advanced Topics in Data Management
2
Class Info

• Instructor Seung-won Hwang
• Office B2 225/ By appointment
• Class B2 101 Tu 1900-
• URL http//www.postech.ac.kr/swhwang/adb.html

3
Welcome!

• Grads
• Non-DB grads for insights/collaborations
• DB grads for
• Preliminaries classic DB papers
• reinventing the wheel? solving new problems
  with time-tested ideas
• Undergrads
• Glimpse of post-graduate life (later)
• No pre-requisite

4
Post-grad life

• Read papers
• Write papers
• Review papers
• Present papers
• Repeat above until graduation/retirement

5
Success at Post-grad life

• Reviews 30 Read/Review
• Projects 30 Write
• Presentation 30 Present
• Class Participation 10
• DONT PANIC. The goal is exposing you to these
  components, rather than judging you by it.

6
Presentation

• You will be assigned as a moderator in 3 classes
  in ENGLISH
• Responsibility
• 50 min paper presentation QA (10)
• Collecting reviews and posting the best review

7
Successful Presentation

• a2b2c2

• High-level intuition
• Illustration example

8
Review

• Summarization of paper in your own wording (Dont
  copy and paste abs/conc or related sections of
  other papers)
• Constructive criticism (what is strong/weak, what
  can be done)
• Questions
• Good for critical reading/writing
• Instructionhttp//www.cs.uml.edu/gary/classes/91
  _421/Misc/Writing_Paper_Reviews.pdf
• Samplehttp//www.cs.uml.edu/cchen/574-S07/sample
  _review.html
• (links available from class page)

9
Review Assignment

• A review (of the paper of your choice) for every
  four papers.
• Send the review to the moderator
• Moderator identifies the best review and posts it
  (If none is sent, moderator writes a review for
  posting)
• Moderator answers/discusses the questions during
  class
• You are strongly encouraged to send more than one
  review for your own exercise (and for higher
  chance to be selected)
• Dont spend too much time focusing solely on
  technical details. Get the big picture
• Advice
• http//www2.cs.uregina.ca/pwlfong/CS499/reading-p
  aper.pdf

10
Project (Success Story)

• Approximate DBIR Query Answering using Wavelets
• Skyline Ranking for Uncertain Data with Maybe
  Annotations
• Minimizing User Interaction for Categorical
  Skyline

11
Project Guideline

• My own experience
• Mostly solo with advisor
• Met ML expert at class, ML front-end for query
  learning query processing ICDE05
• Found out labmate is working on similar paper
  combined into SIGMOD06
• Diverse backgrounds/specialty helps
• Again, DONT PANIC- You wont be alone and will
  be guided by brainstorming sessions, instructors,
  etc.
• Great chance to get a conference paper done

12
Ice Breaker

• Brief self-introduction
• Memory of this summer

13
Your root as a DB researcher

• How many Turing award winners in our field?

14
Knowledge in Creation Database Turing Awards?
1988 Ivan Sutherland 1989 William (Velvel)
Kahan 1990 Fernando J. Corbato' 1991 Robin
Milner 1992 Butler W. Lampson 1993 Juris
Hartmanis 1993 Richard E. Stearns 1994 Edward
Feigenbaum 1994 Raj Reddy 1995 Manuel Blum
1996 Amir Pnueli 1997 Douglas Engelbart 1998
James Gray 1999 Frederick P. Brooks, Jr. 2000
Andrew Chi-Chih Yao

• 1966 A.J. Perlis 1967 Maurice V. Wilkes 1968
  Richard Hamming 1969 Marvin Minsky 1970 J.H.
  Wilkinson 1971 John McCarthy 1972 E.W.
  Dijkstra 1973 Charles W. Bachman 1974 Donald
  E. Knuth 1975 Allen Newell 1975 Herbert A.
  Simon 1976 Michael O. Rabin
• 1977 John Backus

1978 Robert W. Floyd 1979 Kenneth E. Iverson
1980 C. Antony R. Hoare 1981 Edgar F. Codd
1982 Stephen A. Cook 1983 Ken Thompson 1983
Dennis M. Ritchie 1984 Niklaus Wirth 1985
Richard M. Karp 1986 John Hopcroft 1986
Robert Tarjan 1987 John Cocke
15
Database Turing Awards

• 1973 Charles W. Bachman
• 1981 Edgar F. Codd
• 1998 James Gray
• Whos who?
• What have they contributed?
• (stay tuned in this class)
• And we certainly need more!
• (work hard and we are counting on YOU!)

16
System R

• System R 1974-1978
• IBM San Jose Labs, lots of PhD researchers
• Gray coming from OS, first CS PhD of Berkeley
• ACM SIGMOD Innovation Award 1992
• Turing Award 1998

17
INGRES

• INGRES 1973-1977
• U.C. Berkeley faculty graduate students
• Mike Stonebraker (then an asst. prof)
• ACM SIGMOD Innovation Award 1991
• ? Postgres ? PostgreSQL

18
Joint ACM Software System Award 1988

• System R/INGRES
• Citation
• The INGRES and System R systems demonstrated
  that a practical and efficient database
  management system (DBMS) could be implemented
  based on the relational data model.

http//awards.acm.org/software_system/
19
Contributions of System R??
20
Contributions of System R?

• Bringing theory to practice
• Nice theory implemented into practical system
• High level query language (SQL)
• Codds relational algebra/calculus implemented
• Declarative language what (not how)
• Small set of well-understood operators
  (selection, projection, joins,)
• Cost-based optimization
• Enabling technology for SQL
• select people where job programmer and city
  champaign
• path 1 job index --gt check city path 2 city
  index --gt check job

21
Cost-based Optimization

• Cost estimation based on
• index selectivity
• job (programmer) more selective or city
  (champaign)?
• This got me a PhD (old idea, new problem ? theme
  of this class)
• Ideas on views, transactions, ..

22
Thanks