Question

1-Why is a device such as a tape-drive a poor secondary storage device for Random/Direct access to a file?

2-. Describe 3 different approaches that can be implemented with respect to file system mounting.?

3-The concept of a global clock is NOT realistic – why?

Answer 1

1-Why is a device such as a tape-drive a poor secondary storage device for Random/Direct access to a file?
Ans)
Tape-drive is a poor secondary storage device for random/direct access to a file because of the following reasons.

a) Arrival at the desired storage location may be preceded by sequencing through other locations
b) Data can be retreieved only in the sequence in which it is stored
c) Access time varies according to the storage location of information being accessed
d) Suitable for sequential processing applications where most of the data records need to be processed one after another

2-Describe 3 different approaches that can be implemented with respect to file system mounting.?
Ans)
A file system has to be mounted before it can be used by operating system and is available to users. The mount process makes set of files and directories within a disk partition available within the file name space. There are different approaches for mounting file systems. Following are some of them.

a) Microsoft's FAT file system
The FAT approach is used by Microsoft's FAT12, FAT16, and FAT32 file systems. These were used since the days of first MSDOS floppy-disk based file systems (FAT12) and are still supported through today's systems. All of these file systems rely on clusters. The number after the FAT refers to size of the cluster pointers.

FAT16 used 16-bit cluster pointers and supported up to 64 sectors per cluster, offering a maximum file system size (with 512 byte sectors) of 2 GB. FAT32, introduced with Windows 95, increased the cluster pointers to 32-bit values. With up to 64 sectors per cluster, it now supported a maximum file system size of 8 TB and a maximum file size of 4 GB. In all of these systems, file metadata was stored in the directory file.

b) Unix file system
the indexing approach is the method used in Unix file systems (UFS). This was created in Bell labs unix version and is used in AT&T's
unix version as well. The Unix File System's inode comprises 10 direct blocks, one indirect block, one double indirect block, and one triple indirect block. It actually uses 1024-byte blocks and 32-bit block pointers.

c) Virtual File system interface (VFS)
Sun introduced another layer of abstraction called Virtual File System (VFS) Interface to its own version of UNIX. It is been adopted in many other POSIX like systems. The idea behind VFS is to provide object-oriented approach to file system. System calls related to file system interact with VFS directly. The VFS actually provide high level abstraction and is a generic interface for managing files and directories. It interacts with one or more file system modules that implement file system code and is unique to each file system. Each file system module expose common functional interface. File system module interacts with buffer cache and block device drivers to store and retrieve data from actual devices.

3-The concept of a global clock is NOT realistic – why?

Ans)
There are many clocks in distributed system as there are many systems. The clocks are coordinated to keep them consistent but no one clock has the exact time. Even if clocks were in sync the individual clocks on each component may run at a different rate leading them to out of sync only after one local clock cycle. Time is known only with a given precision. At frequent intervals, a clock may synchronize with other trusted clock. But clocks are not precisely the same because of time lapse due to transmission and execution.