Distributed file systems.
Distributed file systems launched in the 1970s, the possibility of merging into
a single computer network has revolutionized the computer industry. This
possibility was first to the division accessing files between computers. The
initial achievements in this area have been limited to copy entire files from
one machine to another. An example is the program UNIX-to-UNIX copy (uucp) and
File Transfer Protocol (ftp). But they did not allow even close to access to
files on the remote machine, in its capacity reminiscent access to the files on
the disk. Only in the mid-1980s, several distributed file systems, which provide
transparent access to network files. These were the Network File System (NFS),
Sun Microsystems (1985), Remote File Sharing system (RFS) AT & T (1986) and the
Andrew File System (AFS), Carnegie Mellon University (1995). These three systems
sharply differed on the objectives of the development, architecture and
semantics, but they are all trying to solve the same basic problem. The RFS is
in almost all systems based on UNIX System V. Develop ASF moved Transarc
Corporation, in which it was developed and turned into a Distributed File System
(DFS), a distributed computing environment komponennt DSE (Distributed Computing
Environment) Open Software Foundation. But most have been NFS, which is
supported on all UNIX and many did not UNIX systems.
Common characteristics of distributed file systems.
traditional centralized file system allows multiple users working on one system
to share access to files stored locally on your machine. Distributed file system
expands the capabilities, allowing shared access to files to users on different
machines that form through the network. The distributed file system based on
client / server model. In this case, the client is a machine that appeals to a
file, and under-server machine that stores files and providing access to them.
Some systems require that clients and servers have different machines, while
others admit to one machine and worked as a customer and as a server. It is
important to note the distinction between distributed file systems and
distributed operating systems. Distributed operating system such as V or Amoeba,
the user looks for a centralized operating system, but also works on several
machines. It can be a file system, shared by all machines system. In contrast, a
distributed file system is a software layer, which manages the connection
between traditional operating systems and file systems. This layer of software
integrates with the operating system maschin-hostov network and provides
services for distributed file access to systems that are central nucleus.
Distributed file systems have a number of important properties. Every system may
have all or part of these properties. This is precisely the basis for the
comparison of different architectures among themselves. Network transparency -
Clients should be able to access files using the same operation, and that access
to local files. Transparency of-Filename should not determine its location in
the network. Independence of-filename can be changed with the change of its
physical metoraspolozheniya. Mobility user-Users should be able to access shared
files from any network node. crash - The system should continue to operate
under separate component failure (server or network segment). However, this can
lead to performance degradation or eliminate access to some of the file system.
Scalability - The system must be able to scale in the event of an increase in
workload. Furthermore, it should be possible to gradually increase the system by
adding individual components. Mobility-files should be possible to move files
from one location to another for a working system.
|