Suppose a RAID 4 disk system has 4 disks: D1 to D4, where D1 is the redundant disk and the others are data disks.  Suppose the bit value at a fixed position in D1 is 1, in D3 is 0, and in D4 is 0. What is the bit value at the same position on D2?    

We have a disk that has 200 tracks and uses the elavator disk scheduling to serve requests for data on the disk tracks. Suppose at time 0, the scheduler has received requests for data on the following list/queue of tracks: 36, 69, 167, 76, 42, 51, 126, 12, and 199, in that order. If the r/w head is currently on track 100 and is moving to track 200, the request for data on which track will be served AFTER serving the one on track 76? Your answer is one of the tracks listed above. 

Which of the following is CORRECT about comparing RAID 5 and RAID 6?

Suppose we use a disk scheduling algorithm, called Shortest Seek Time First (SSTF), which always serves the closest track first. (For example, if the r/w head is currently on track 53 and there are requests for data on tracks 57 and 46, it first serves the data block on track 57 and then the one on track 46.) Suppose the scheduler has requests for data on tracks: 14, 37, 65, 67, 98, 122, and124. If the r/w head is currently on track 53, what is the total number of tracks moved over to serve all these requests? For our example above, the answer is 15, obtained by adding 4+11 (4 moves from 53 to 57 and then 11 moves to 46).      

True/False: It is possible to have a linear hash structure with n>1 primary buckets and about n overflow  buckets in which all but one primary buckets are empty and almost all overflow buckets are full.  

Which of the following statements is CORRECT about finite bit vectors/strings?

Consider a grid file on relation R(A,B) with both attributes in the range -100 to +100, such the the grid lines in both dimensions are located at positions that are multiples of 20. Suppose we want to find the points in R that are closest to the point P(40,60). If the query returns the point Q(50,50), how many other buckets we should look into to verify that Q is indeed the closest point to P? 

Consider a linear hash table in which each bucket can hold upto 3 records. If the utilization ratio is 70%, what is the maximum number of records stored in the hash table if the number of primary buckets used is 15 and number of overflow buckets used is 2?

Of the following 4 index structures, how many can be used to support both partial and range queries?

• B⁺-trees
• Grid files
• Partitioned hashing
• KD-trees 

We have built a partitioned hash index with 5 bits on a table S(A,B,C). Suppose the first three (left-most) bits are for attribute A and the last two (right-most) bits are for attribute B. How many primary buckets will be searched to find records in S with B=25?