A computer network
server can be configured reliable
and efficient according to the needs
of a business. If the business is
of a critical nature like banks, consulates,
police department or corporate companies,
the cost on the reliability of the
server increases. There are a large
number of the objects involved in
a computer network such as hubs, routers,
Gateway, bridges, ISDN adapters, Modems,
transceivers, firewalls and many others
and all these devices by some means
depends on the server to share data
and resources. There are a number
of the subsystems involved in the
reliability of a server. Following
are the tips and recommendations in
making a secure, high performance
server according to your business
needs.
1. Hard Disk
Subsystem Reliability
The main server component
that is prone to failure is the hard
disk that stores the data of the server.
Hard disk is a mechanical component
in a server. While selecting the disk
drives, you should consider the S.M.A.R.T
technology (Self monitoring and Reporting
technology). This technology allows
the monitoring of the disk drives
parameters and is able to predict
the failure of the disk drives. Most
server vendors provide the hot swap
disk drives this means that they can
be replaced while the server is in
working mode. To provide fault tolerance,
implement RAID (redundant array of
the inexpensive disks). RAID is implemented
by installing RAID disk controller
in the server. RAID 5 provides fault
tolerance by adding the parity information
of the data as it is written on the
disk drives. This parity information
can be used to regenerate the data
in case of disk failure. Minimum 3
disk drives are required to implement
a RAID 5 (striped set with parity)
and the maximum numbers of the disks
involved in the RAID5 are 32.
2. Server Processor
Most of the servers
use the Pentium processors that are
very reliable in performance and processing
speed. Processor or CPU is a heart
of any server and if it fails to work,
the server becomes non functional.
Trend is growing to use a dual-processor
server or even multiple processor
servers. If this scenario, if one
processor fails to work, there is
not effect on the other processors
in a server. So it is a best decision
of the network administrators to use
the multiprocessor server.
3. Network
Subsystem
The attachment of the
server to the network is carried out
by the NIC adapter. If a NIC of the
server fails to work, the connection
of the server with the other computers
in a network also fails. Many servers
support redundant NICs that can be
implemented in the fail over or load
balancing situations. In the NIC fail
over situations one NIC is used to
communicate with the server and the
other NIC is held reserve until the
other NIC fails and the second NIC
is automatically activated. In the
NIC load balancing situation both
the NICs are activated so improve
the reliability and performance of
the server communication. A PCI hot
swap slot allows replacing the faulty
NIC adapter without shutting down
the server.
4. Clustering
Redundancy is achieved
by clustering. In a clustered environment,
two or more server operates as a single
server and the failure of the one
server, does not affect the working
of the other server. There are many
advantages of the clustered environment
such as reliability, load balancing
and high performance of the computing
environment.
5. Memory Performance
The memory technology
has been improved and it fails to
work very seldom. Purchasing a memory
for a server that supports ECC (Error
correcting code) is a best decision
of the system ornetwork administrators.
ECC technology is capable of detecting
and fixing the single bit error and
reporting the multiple bit memory
errors. A server needs a high amount
of memory to perform its function
smoothly. If a memory is low in a
server, it can results in slow boot
up, slow response time and halting
the system so memory is a vital for
a computing server.
6. Power Supply
Usage
Many computers come
with the multiple power supplies and
these multiple power supplies 2 or
more share the electric load of the
computer server. The main advantage
of the multiple power supplies is
that the failure of the one power
supply does not cause the server to
shut down because other power supply
takes all the electric loads and so
server works normally without any
problem. Like the other hot swappable
hardware in the server, the multiple
power supply system is also hot swappable.
This means the faulty power supply
can be replaced with the new one,
while the server is running.
7. UPS ( Uninterruptible
Power Supply) Function
A UPS or uninterruptible
power supply system plays very vital
role in case of the power failure.
UPS normally provides power back for
3-4 hours and during this time, server
can easily be setup to save the data
and all the network users can also
access the store their important data
on the server. The shutdown process
of the server is controlled by software
that is installed on the server. UPS
is very important in a network environment
and if there are a number of the UPS
in a network, they can share the load
and provide constant power supply
to the server.
8. Emergency
Generator Working
When a full time operation
of the server is required such as
hosting servers, web servers, email
servers, data server, and emergency
generator is a necessary. The emergency
generators are usually diesel powers
and are developed by a number of the
world’s renowned companies like
Jubaili Brothers, Cummins. These emergency
power generators can be set to automatically
switch on in case of power failure
or they can be used after the UPS
is fully used.
9. Cooling
Fans Function
There is a large amount
of the hardware installed in inside
the CPU so there must be cooling system
inside the CPU. The cooling of the
server is achieved by the cooling
fans, which are installed inside the
CPU. These cooling fans pull the cool
air inside the CPU and throw away
the heated air from inside the CPU.
The failure of the cooling fans in
the server leads to the automatic
shut down of the server due to high
buildup heat in the server. Many vendors
are providing the hot swappable cooling
fans.
10. Computer
Monitoring Hardware
There is a specially
designed adaptor available for monitoring
the performance of the different components
of the server. The server can be monitored
locally or remotely via a modem by
this adapter. In very critical network
environments, these types of adapters
are usually used in the server to
put a check on the performance of
the different networking devices.
11. Computer
Server Software
Software installed
on the server should be very reliable
and virus free. A good anti virus
software like Norton Anti Virus, Trend
Micro, McAfee Anti virus, Panda Anti
virus or any other anti virus should
be installed and updated. Additionally,
a good firewall, anti spyware, malware,
adware should also be installed to
keep the server virus or spyware free.
Good backup software should also be
installed on the server in case of
any emergency. Any unwanted software,
media players, codec, flash games
or emails can provide a room for the
viruses from the internet to enter
in your network. No unauthorized software
applications or hardware should be
installed in the server. If a virus
enters in the server by any means,
it is very easy for the viruses to
spread all over the network and if
they are of severe nature, they can
even destroy all your critical data.
The above mentioned recommendations
for a designing a good server can
play a key role in designing a secure,
reliable and high performing server
for your computer network.
12. Data Server
Location
A server should be
placed in a room that is capable of
providing the cooling and dust free
environment to the server. It is very
advisable that the server room should
be very secure and only the authorized
persons like (network administrators
or other users) be allowed to enter
in the server room.