What is it?
A file server is a special purpose computer used in a network environment to
provide file-related services to users. A file server can be a mainframe or
minicomputer, but within our context, generally means a PowerPC or
Intel-based CPU. In comparison to a general-purpose workstation, a file
server generally has increased disk capacity, increased memory, and may have
a faster bus architecture than a normal workstation. Typically, the video
subsystem on a file server is less important than on a workstation, since it used
only to provide access to operating system utilities, and management functions
for the server. A server must obviously also always include at least one
network adapter.
At the high-end, file servers may be designed with additional features that
enhance their speed and reliability. These reliability features generally involve
redundancy, a term used to describe design considerations that allow a
portion of the server to fail, without causing a complete shutdown of the server.
Redundancy may apply to the disk subsystem, in that a server may have
duplicate disk controllers and disk drives, that store multiple copies of data, in
case one of the drives or controllers should fail. They may include dual power
supplies, and may, at the extreme, include multiple processor units, each
sharing the computing load, so that a failure in one processor will only slightly
degrade the performance of the server. A server may also include multiple
LAN adapters, that can provide both redundancy, and the ability to route
between separate network segments. Other considerations in server might
include the ability to change components in the server while it is powered-up
and running, known as hot-swap capability.
What does it do?
A file server is primarily designed to provide a central, secure repository for
user programs and data files. It may also provide access to shared resources,
such as network printers and communications services. These capabilities are
enabled by loading and running the network operating system (NOS) on the
server, along with the appropriate client software on the attached workstations.
A server can be either dedicated or non-dedicated, depending on the NOS
involved. For example, Netware versions 3.x and above require dedicated
servers, which means that while acting as a server, the CPU is unable to
perform other functions. This limitation is somewhat eased by the ability to load
NLMs (Netware loadable modules) which allow the server to provide some
auxiliary processing in addition to running the file services. In other NOSs, such
as AppleShare, Windows NT, or OS/2 LAN Server, the server operating
system runs within a multi-processing environment, so the server can run other
processes concurrently. In other words, it is non-dedicated.
How is it used?
A server provides the foundation for most scaleable networks used in a
corporate environment. Although there are peer-to-peer networks that allow
users to share files located on their local disks, a network of any significant size
is designed so that a server provides centralized services to a group of client
workstations. A server and its associated network operating system provides:
- a way to control access of users to files, directories, applications, and
services such as printing;
- a central point for upgrading applications, and backing up data;
- a foundation for sharing applications and data between users;
- a way to enable group productivity applications such as electronic mail
or groupware;
- a way to monitor the performance of the network, and the activities of
the users.
Where do I get more information?
Contact your local CBV Office.
|
