Very small scale companies generally do not have a server. Even if they use some of the PC, the only peer to peer network PC. When demand started to increase, peer to peer network is no longer adequate. Needed a network with the addition of switches and computer servers.
When the server must choose, many people are surprised by the price of a server. This is indeed fair, because the price of a server can more than double a desktop PC. What is the difference between a normal desktop PC with a PC specially designed for server? Perhaps a simple analogy that can be used is the difference between a car and family car truck. An family can only be used for hauling goods, but when the tasks that must be done is to carry goods, the trucks are more effective and efficient. In brief, there are some important things that must be taken to decide between the PC desktop with a PC server:
On the PC desktop, the performance of the processor is generally seen from the speed (MHz). In the PC server, in addition to speed, cache role (even though the desktop PC is also important), but on a PC server is far more important. Not only that, the processor for servers must have criteria, reliability, quality, and are validated by the various operating system. If the note, the processor for servers have L2 and L3 cache, which ranges twice the size of L2 and L3 cache on the processor for the desktop. For example, the latest Xeon processors have 512KB of L2 cache, and L3 cache of 4 MB, while the Pentium 4 processor Extreme Edition has L2 cache 512KB cache and L3 2 MB. Processor server of course, requires a special chip that is designed to support.
On the PC desktop maximum memory capacity of 4GB supported only, while the PC server (32-bit) supports a maximum memory up to 16 GB.
The main circuit board on the server has many features that are far more complete and reliability that is far better than the desktop motherboard. Server motherboard supports automatic monitoring feature, which allows engineers treat the server easily and quickly analyze trouble. Motherboard also has chip that works to save the log all activities, including when someone opened the front door or the side casing. Supported for several processors designed for random, so you can boot from one processor if the other failure. Processor can take power from more than one source, thereby reducing the power on the server. On the server motherboard also supports the use of more than one processor.
The server chassis is designed specifically to support the motherboard that has specifications. This chassis has a security and monitoring features built-in help to bring the reliability and security maximum. Storage capacity can be improved quickly, even without the need to turn off the server. In some chassis provided indicators to indicate the condition of the components in that server.
PC desktop generally has only one network connection, while the server uses two network connections: one main connection, a backup connection. In addition to a backup, connect the two also help increase throughput.
Different from the PC desktop, which generally use the ATA hard disk (serial ATA or parallel ATA), PC servers generally use a SCSI hard drives that have a higher lap speeds, greater memory, Throughput higher and reliabilities better.
Raid is a feature to increase protection against readability and data on the server. PC servers that use hard drives configured based standard that allows the RAID servers continue to work even though there is a problem in one of the RAID hard disk. Almost a server has two hard disks in which a hard drive as the main hard disk, hard disk and the second as a backup hard drive. Data on the main hard disk completely copied to the hard disk backup automatically. If the main hard drive suddenly broken, hard drive backup can take over without losing a bit of data. To make PC servers needed to support RAID controller installed on the motherboard.
In addition to simplify the management of servers, software to monitor the operation of this work a number of important device on the server, so it can know the problems that may occur. Thus, the user network can overcome the problem.