Assignment 4: Replication, Design, Query Processing
Start date 24 April, due beginning of class 1 May.
Complete the following exercises:
-
Book exercise 8.6.
(If you like the challenge, you can take a look at 8.7/8.8,
but I didn't want to overload next week.)
-
Book exercise 9.2.
-
Book exercise 9.3.
-
Repeat exercise 9.2/9.3, but this time assume a
multiprocessor system. Describe what join
algorithm you would use
(see slides 13.32-13.35) and why.
-
Assume the
Dataspaces
concept catches on. In relational terms, assume that these dataspaces
support queries with select, project, and equijoin. You decide that
turning dataspaces into fully functional distributed databases is
an exciting research area.
-
Describe one
low-hanging fruit
: some functionality not
covered in Halevy's concept that you think would be relatively
easy to provide (perhaps challenging enough to get a paper
out of, perhaps not even that.) Briefly sketch how you think
it might be solved (e.g., we can apply ... from distributed
databases directly
.)
-
Describe a problem you think will be really hard - a feature
we expect of databases that will be very difficult to support
in the dataspace framework. Give a brief sketch of
why this is hard.
I'm looking for about a paragraph for each part - just
a high level description.
Turning in the assignment
Electronic submission preferred, upload text or PDF (or
check with me for other formats) to
Blackboard.
Hard copy is acceptable as well, please hand in at the beginning of class.