Servers
Real Servers Don’t
Eat Quiche
While essentially any desktop PC with
network access can operate as a file server or networked storage serving
data, files, or applications, true servers are constructed with power,
durability and consistency in mind. Our objective with this manual is
not to turn CPA’s into IT specialists but rather to educate CPA’s on
technology topics they need to be familiar with. The following are some
basic points we think everyone should know about servers.
Basically a
computer on a network that is dedicated to a particular purpose like web
serving of file serving is a server. A network’s server is responsible
for holding the files and managing the processes that enable everyone in
the office to logon and use the network. Similarly, a file server holds
the central computer files or databases for other computers to access.
Today’s dedicated
servers feature many of the same components as modern desktop PCs; such
as a motherboard, RAM, a CPU, and hard drives. However, they convey far
stronger performance and reliability based on their beefier parts.
Whereas all desktop PCs (32-bit) are limited to 2-4GBs of RAM, many
servers can accommodate 8GBs or more. Even LAN connections are faster.
Because a server provides multiple PCs and users with resources it must
make that information freely available and fast. Most serves today are
equipped with Gigabit Ethernet, giving approximately 10 times the
throughput speed of an ordinary 10/100BaseT NIC. The CPUs for servers
are also engineered differently. Server CPUs are typically 64-bit
processors and thus have long required 64-bit operating systems and
applications. However, newer server CPU’s like AMD’s Opteron offer the
ability to slowly migrate to 64-bit computing as application development
evolves. These new 64-bit processors and their desktop relatives can run
32-bit programs with increased efficiency while maintaining the raw
64-bit power of cutting edge computers.
Lastly, server hard drives can be of
sturdier construction and are usually designed with more fail-safes than
their desktop equivalents. Many servers are setup in a Redundant Array
of Independent Disks or RAID configuration. There are countless
configurations for a RAID array all specializing in different methods of
data protection or performance enhancement. The simplest setup is a RAID
1, or mirroring array. RAID 1 duplicates the contents of your primary
drive onto an auxiliary drive to guard against physical drive failure.
If one or the other crashes, you have a perfect backup. However, don’t
be confused about the amount of storage available. If a server or
desktop is configured in a RAID 1 array the storage capacity will only
equal half of the combined drive size because one is an exact copy of
the other.
The most current trend in servers today is
blade servers. These nearly self-contained computers are constructed
entirely on a thin chassis base which can then be slid into a
rack-mounted interface along side countless other “blades” creating one
giant uber-server as shown in the picture above. The rack-mounted
servers do require a degree of space management and planning as well as
some environmental control. Blades require adequate ducting and
ventilation as the exhaust heat from one server blade can conceivably be
transferred and increased all the way through the server rack to the
last blade, dangerously reducing its life due to constant heat and
pressure.
- END -
