[Announcements | Schedule and Readings | Assignments and Quizzes | Syllabus/Policies ]
Reading Questions, Lecture 22 (11/29)
The CONTROL project investigated online query processing, which is yet another
"continuous query processing"-like system.
This work immediately preceded the Eddies paper that
we read earlier and is focused on producing a continuously updated set of answers
to a long running query in a traditional relational database. Answers include
some notion of uncertainty that decreases as the query runs.
As you read the paper, consider the following questions:
- When is the CONTROL approach appropriate? What kinds of users would find it useful?
- Why does the CONTROL system need to randomly sample the data? How does it do this?
- How does the CONTROL system measure uncertainty? Is this metric appropriate? Are the assumptions used to measure uncertainty reasonable?
- Are any of the CONTROL techniques useful in the streaming systems we have studied?
Do streaming systems obviate the need for CONTROL?
Samuel Madden (madden at csail dot mit dot edu)
Last modified: Thu Dec 2 16:01:19 EST 2004