Files and file systems
The first computers each application, which was required to store data in
external memory, to determine the location of each portion of data on magnetic
tape or drum and served exchanges between operational and external memory with
the help of computer commands or call the relevant programmes operating system.
Such an operation does not allow or significantly hampered maintenance on one of
several foreign media archives long stored information. In addition, each
application had to deal with naming of the parts and their structure in external
memory.
Revolutionary step was a shift to centralized management systems files. In terms
of application file - a region known as external memory, which can be saved and
from which you can read the data. The rules for naming files, way to access data
stored in a file, and the structure of the data depends on the specific
management system files and, possibly, of the type of file. File management
system assumes the distribution of external memory, display file names to the
appropriate address external memory and access to data.
In all file systems explicitly or implicitly allocated some basic level,
providing work with files representing a set of directly addressed in the
address space file blocks. These logical blocks multiple of the amount of
physical disk block, and usually chosen equal virtual memory page, which is
supported by computer equipment in conjunction with the operating system.
In some basic level file systems available to the user, but more often it rose
higher, standard for all users. In practice, have proliferated in two basic
approaches. In the first approach, users represent file as a sequence of records.
Each record - this sequence of bytes permanent or variable size. Entries can be
read or write, or to position consistently file on record with a specified
number. Some file systems allow structured records in fields and declare some
fields keys record. In these file systems can require sample records from its
file on a given key. Naturally, in this case supports a file system in the same
(or another - Service) core file additional office invisible user data
structures. Common ways of organizing key files are based on hashing and
Technology B-trees.
The second approach is that any file submitted as a sequence of bytes. Because
you can read the file specified number of bytes from its beginning or making his
pre-positioning of bytes with the specified number. You can also write specified
number of bytes at the end of file, or making his pre-positioning.
Modern file systems support multilevel naming files from the maintenance of
additional external memory files with a special structure - catalogues. Each
directory contains the names of directories and / or files contained in this
directory. Thus, the full name of the file consists of a list of names of
directories, plus the file name in the directory containing this file directly.
The difference between the way for naming files in different file systems is how
to begin this chain of names.
In many file management systems require that each archive files - a tree full of
manuals, are entirely located in one package or logical disk drive - the section
of physical disk package, submitted by means of operating system as a separate
drive. In this case, the full file name starts with the name of the disk device.
This method is used in naming the company file systems DEC. Very close to naming
files in this way are file systems of personal computers. This organization is
called the maintenance filenaming isolated file systems.
Another option is the last of the totality of directories and files as a single
tree. Full name of the file name starts with the root directory and the user is
not obliged to care about the installation disk device on any specific discs.
The very system by searching the file by name, asks to install the needed discs.
Such a file system called a completely centralized.
Centralized file system is largely isolated systems easier, as they file
management system assumes more routine work. But in such systems has had a
substantial problem, if someone is required to defer podderevo file system to
another computer installation. A compromise solution used in file systems, OS
UNIX. At the basic level in these file systems are supported isolated archives
files. One of these archives declared root disk. After running the system could
"mount a" root and a number of separate file systems in a common file system.
Since the file systems are a common data repository, generally belonging to
different users, files management system should provide authorization to access
files. Such an approach is that in relation to each registered user of the
computer system for each existing file indicates the action that he allowed or
prohibited. In most modern systems of government files every registered user
corresponds to a pair of integral identifiers - identifier group to which the
user ID and his own group. In this case, every time a file is kept full ID,
which created the file, noting what actions with the file may make himself what
actions with the file available to other users of the same group, and that could
do with the file, users of other groups. This information is compact, when
checking is not required large quantities of action and a way to control access
sufficient in most cases.
If your operating system supports multi-mode, it is real situation where two or
more users simultaneously trying to work with the same file. If these people are
going to only read the file, then no problem arises. But if one of them will
change the file, then to perform the work of this group requires mutual
synchronization of the operating system.
Consider the possible areas of application file systems. First and foremost, of
course, the files used for storing text data: documents, texts of programmes and
the like. Such files typically are formed and modified with text editors. The
structure is quite simple text files - is a sequence of records containing lines
of text or sequence of bytes, among which there are special characters, for
example, the end of the line.
Files with the texts of programs are used as input texts compilers, which in
turn create files containing object modules. In terms of file system, object
files also have a simpler structure - a sequence of records or bytes. The
logical structure of the same object module unknown file system, and supported
programs on the system programming.
Similarly, the case files compiled editors ties and containing images of a
running program. The logical structure of such files remain known only to editor
ties and loader - program operating system. Roughly the same situation with
files containing graphic and audio information.
Thus, file systems typically provide storage poorly structured information,
leaving the structure further applications. In the cases listed above, this is a
positive feature, because it allows when developing software, using simple and
relatively cheap means of a file system, to organize the structure of data
storage, the most relevant specifics of Applied area.
|