10.12 RAID systems typically allow you
to replace failed disks without stopping access to the system. Thus, the data
in the failed disk must be rebuilt and written to the replacement disk while
the system is in operation. Which of the RAID levels yields the least amount of
interference between the rebuild and ongoing disk accesses? Explain your
In my judgement, RAID Level 1 suits best for
the given situation. It yields the least amount of interference between the
rebuild and ongoing disk accesses. In rebuilding period, RAID Level 1 copies
only data from failed mirror disk. But in all other RAID levels, RAID copies
every content available in the disk.
10.14 In the variable-length record
representation, a null bitmap is used to indicate if an attribute has the null
a. For variable length fields, if the
value is null, what would be stored in the offset and length fields?
b. In some applications, tuples have a
very large number of attributes, most of which are null. Can you modify the
record representation such that the only overhead for a null attribute is the
single bit in the null bitmap?
length record, null bitmap is used to represent the attributes of record which
have null values. In null bit map, the offset value is set to 1 and the length is
Yes, by storing
the null bitmap in the beginning of the record. It is useful for tuples having large
number of attributes which are mostly null. This record representation would
save storage space and extra work cost for retrieving attributes of the record.
10.18 In the sequential file
organization, why is an overflow block used even if there is, at the moment,
only one overflow record?
the sequential file organization, the overflow block is used to insert records
if no free space available to insert in sequential position. Then pointers
adjust accordingly. Block is the smallest unit in the disk. Based on performance
perspective, storing record in overflow block is better than inserting multiple
records in block. The free space in the block may use for future insertions.
If you have data that should not be lost on disk failure, and the data are write
intensive, how would you store the data?
is evolved to control disk failure. Redundancy means storing extra information
which normally not useful, but it can be used to rebuild the lost information
in disk failure. “Mirroring” is a redundancy technique which duplicates every
disk. If one disk fails, we can read the
data from another disk. So, by using mirroring we can avoid data loss on disk
10.21 In earlier generation disks the
number of sectors per track was the same across all tracks. Current generation
disks have more sectors per track on outer tracks, and fewer sectors per track
on inner tracks (since they are shorter in length). What is the effect of such
a change on each of the three main indicators of disk speed?
three main indicators of disk speed are seek time, rotational latency and data
transfer rate. First, to access the data of a required sector, the arm needs to
move over the correct track first and then it needs to move to the given sector.
The time taken for repositioning the arm is called seek time. In modern disks with
more sectors per track have faster seek times than the disks which have less
sectors in older days. In earlier generation disks, the area of sector is more,
so moving arm from one sector to another sector takes more time. Secondly, the
time used to access the sector under arm is called rotational latency. Because
of less sector area in modern disks, rotational latency also decreases.
Finally, the retrieving rate of data from or to stored disk is called data
transfer rate. Current disk systems are supporting maximum data transfer rates
per second. Inner tracks data transfer rate is less than the outer tracks
because of fewer sectors in the inner tracks. So, for this reason older generation
disks have more data transfer rate.