Question 1:
a) What are the communication requirements for building distributed
database systems? (6)
b) What facilities are needed for efficient distributed transaction
processing
functions? (7)
c) What do we need to do for communication facility to build a large
scale
distributed system (say 1,000 site system). (7)
Question 2:
a) Quote some numbers for the latency of communication or round trip
times
for all types of networks across all distribution of physical sites (interprocess,
LAN, MAN, WAN in US or around the world). (4)
b) How can we measure the delays and round trip times between multiple
sites
and how can we improve them? (8)
c) What do we need in transaction processing if we can not improve
the
communication behavior any further and must deal with the way it exists?
(8)
Question 3: How would you change the design and algorithms for distributed
database
processing if the communication delay was three time higher than normal?
(20)
Question 4: How would you change the design and algorithms for distributed
database
processing if the communication delay was almost zero? (20)
Question 5: How would you change the design and algorithms for distributed
database
processing if there were multiple and frequent failures (like sites failing
and
recovering, network partitions, variable communication delays, many lost
messages) ? (20)
Question 6: Consider applications such as Digital Libraries, Electronic
Commerce, Video
Conferencing.
a) What type of communication behavior (in terms of round trip times,
reliability)
are required in such applications? (7)
b) How can we deal with the bandwidth and network capacity to realize
such
applications and make them a success. (7)
c) What will be the performance evaluation criterion for such systems?
(6)