is used to manage the shared resources of the network.
The hardware may be a PC or a computer designed
specifically to act as a server. A network server is
able to control network traffic as well as the sharing
of other resources, such as application programs, disk
space, data files, and printers. In either case, the
computer normally has a hard disk and the software
needed to run the network system. There are several
different types of servers, and each has a particular
function. In newer systems, some separate server
functions are combined into a central file server. The
servers we will look at are the disk server, the file
server, and the print server.
The disk server was the first of the network
operating systems. In the early days of PC networks,
very few computers were equipped with a hard disk.
When the hard disk became affordable, manufacturers
were asked to develop a system to allow several users
to share a single hard drive. The earliest disk servers
were multiplexers that polled each connected
computer for requests to write a file on the hard drive
or to retrieve a file from the disk. The multiplexer
then responded accordingly. A major problem with
this process was that it did not allow for any type of
security, data organization, or disk management.
As LAN technology evolved, the development of
the disk server software in the early 1980s addressed
some of these issues. The disk server is a software
routine that was installed on each computer in the
network. The disk server software allowed each PC
to access the shared hard drive as if it were a local
drive. In other words, the computer thought the drive
was installed in the computer, but in reality, the drive
was remotely located on the network.
The disk server also provided for some
information sharing. One purpose of a network is to
allow multiple users access to the same information.
One problem encountered with early disk servers
occurred when two or more users updated the same
file at the same time. When the file was saved by
both users, the updates of one of the users was lost.
A method of preventing this information loss is
file locking. File locking means that when one user
accesses a file, all the other users are prevented from
accessing that file until the first user is finished with
it. As you can see, this method severely limits the
number of users able to access the information.
Another method used to prevent data loss is record
locking. In a data-base environment, many users
could access the same data file, but when a record was
being modified by one user, the other users were
locked out of the record being modified. A data file
can be updated by several users without threatening
the integrity of the data by using this technique.
Although the disk server was used in most LANs
developed before 1985, a major problem still existed
in maintaining data integrity. The two methods
covered in the previous paragraphs provided for data
management, but not for reliable disk management.
A disk drive stores information on the next available
block on the disk. When the disk server was used, it
was not uncommon for two users to try to write data
to the same block at the same time. When this
happened, the second user would overwrite the data
just written by the first user, causing a loss of data.
The development of the file server in 1983 solved all
of the problems encountered with the disk server.
Currently, all local-area networks require some
type of file server. In most cases, the file server is a
dedicated PC or minicomputer. The file server
performs the processing of the network control
software and the central processing and storage point
of the application software and data files of the
network. The file server has a hard disk with a very
large storage capacity.
The file server manages the hard disk and ensures
that multiple requests for the same file do not conflict
with each other. In the disk server environment, each
PC workstation manages its I/O with the disk through
low-level sector calls. In the file server environment,
each workstation communicates with the central disk
through the use of high-level calls to the file server.
A high-level call can be a request to open a particular
file or to store a file, while a low-level call maybe to