Unlike tower servers, offices, rack servers are, which are traditionally installed in 1U, 2U, 3U, 4U or 6U of rack used. The numbers to give names to the Rack Server exactly the number of us, they occupy on the shelves. The 1U server in a single 1U bay, showing two occupy 2U and 4U four, while a rack of standard-size has to 42U:

 

The servers in a 1U format by companies for use in data-hosting and preferred because they are very compact (only 4 4 cm high), are the installation of a large number of servers per rack possible. The main limitations of this format is the limitations on the air (because of the small inner room), inhibit the use of processors with high energy consumption and the need for special radiator and fan use, all running at an extra charge. In addition to the basic components of space to install usually 2 or 4 hard drives of 3 5 “(after the provision of other components) and a single expansion card, horizontal, installed with the help of a riser.

 

Then there is the 2U server. They use “normal” power sources and coolers, and therefore end up being a little cheaper. The larger interior makes the format more suitable for 2U server with two or more processors, or processors to use the high consumption. The amount is not enough expansion cards to install upright, as in tower servers, but you can use a riser (as in the case of 1U), or use half-height cards (the lower panels, half the height of the normal plates) have .

 

Finally, we have the largest servers, the use or 3U 4U. There are 6U server, but they are rare: this format is usually through the disk arrays and blade server chassis used for (see below). With a 3U Server or greater eliminated the problems with the space, so that the expansion cards fit mounted in vertical, and a large number of removable hard drives in bays, but causes the server to more space in the rack, which occupy increases the costs to the Host of the data center, where you would pay an additional fee per used pin.

 

Another format that is becoming increasingly popular, the blade servers (blade comes from the word “Blade” that is, the small format), a brilliant idea further increase the density of servers and allows the sharing of components in the data center , as power supplies and optical disks.

 

The idea is that instead of 10 1U servers, with 10 sources (or 20, if sources were used redundant), 20 network cable (each server typically uses two lines, one for the network and another for the management or redundancy) as well as power lines, KVM, and so on are used, you can use a single chassis, with a corresponding number of blade servers.

 

Each blade is a complete server with its independent processor, memory, network and hard drives. Its small size, the blade servers typically use low-power processors and hard drives of 2 5. At first it was the sharing of processors from Transmeta and VIA, but in the end is almost entirely from Intel and AMD processors, multiple cores are much faster, but still replaced relatively Economics. In the case of hard drives, the discs of 5 second “the favorite for the provision of lower access times (but with respect to rate transfers are losing), in addition to power consumption.

 

Thought not all are equally Blade system created. Therefore, if you are planning to campaign in the blade, which you most likely blade system for your case, you should jump. Check out the following comparisons to help you to compare between the best of them:

 

HP BladeSystem vs. IBM BladeCenter System

 

HP vs Dell Blade system Blade System