1 There are two relations R1(A, B, C) and R2(B, D, E). Which of following techniques CANNOT be used to improve the performance of the natural join of the two relations?
A make use of multiple parallel processes on a multi-core machine to implement the join algorithm
B increase the CPU L2 cache hit ratio of the join algorithm implementation
C partition the relations to different machines by using join attributes and execute the join on many machines in parallel
D when the number of tuples in R2 is large and the number of tuples in R1 is small, build and index on R1 and use index join method
E none of the above
2 Which of the following statements about checkpoints(with log-based recovery schemes) in databases are/is NOT correct?
A checkpoints are used to decrease the time used for failure recovery
B checkpoints are done periodically in database to assure ACID properties of transactions
C periodic checkpoints decreased the number of log entries that need to be redone during recovery
D during checkpointing, updated disk blocks are written to disk
E none of the above
3 There is a huge relation R(A,B,C) of size M and A is the primary key. The tuples in R are sorted by using A and no index is built for R. Each disk block size of b, and each block read takes N seconds. To find a tuple in R with A =17954 the estimated optimized cost is
A M*N/b
B N*log(M/b)
C M*N/(2b)
D (M+N)/b
E log(M*N)/b