What are the different types of servers sold commercially?
There are two main types of servers: the family of RISC processor servers which is the upscale market and the family of x86 architecture servers positioned at the entry level.
RISC processors are adapted for heavy repetitive calculations with their simple design allowing them to be more effective than x86 servers. Marketed by players such as IBM, HP, Bull and Sun, RISC servers see their scope reduced more in favour of entry-level servers, becoming cheaper and more effective. However, they remain the benchmark in the field of high performance computers (HPC), as used in supercomputing research centres.
The family of servers using x86 architecture dominates the market. It refers to the earlier Intel processors, which contrasted with RISC processors voluntarily by offering a complex instruction set, including integrated multimedia functions (audio, video and image). These inexpensive computers were first conquered by the consumer market through the workstations. But in recent years, this processor family is required on the server market through the chips of Intel Xeon and AMD Opteron. In Q4 2004, these servers accounted for 44% of total turnover in the sector.
What do the terms racks, blade or 1U mean?
They designate the organisation and server capacity. A server rack corresponds to a space within the housing to the server on which it is possible to introduce expansion cards to facilitate the evolution of the machine and maintenance. Within a rack, 1U is a standard measure that refers to the usable vertical space. It corresponds to 4.4 cm or 1.75 inches. 16U server may therefore contain more expansion cards and thus extend its computing capacity.
Depending on the task to which the server is intended: file server, mail server, web server, application server or database server, the machine will be well advised to have a strong capacity for expansion to include more computing power.
A server blade comes in the form of an expansion card. But unlike a regular card that includes the processor, it integrates all the elements necessary to a server (disk space, RAM, controllers, input / output, etc). For businesses, these servers offer a genuine interest through saving space and energy. They are often dedicated to simple tasks that can be easily spread, such as web access management, data encryption or job sharing.
Why would someone buy a server rack?
Rack servers were designed to sit comfortably on a shelf and so they created racks for them to go on. Most of them have some sort of speciality, such as being noise proof or ventilated. They are often used to save a lot of space and get rid of a “computer” room by putting all of their servers and hardware into one place. They are also able to handle heat just in case some of the hardware gets hot. It is better to have hot things locked away rather than having them in places that may provide fuel for ignition such as carpets, curtains and wallpaper. They also make moving your company or office to another building a lot easier because you can simply wheel your hardware out within the cabinet.