Parallel & Multi-Core Processing

Blade Servers

Text Box: Blade servers are self-contained computer servers, designed for high density. Whereas a standard rack-mount server can exist with (at least) a power cord and network cable, blade servers have many components removed for space, power and other considerations while still having all the functional components to be considered a computer. Each blade has it own CPU. The work is divided between the processors so blade servers are a form of parallel computing. The chassis, which is the enclosure or case housing the blades, provides services such as power, cooling, networking, various interconnect and management for all of the blades it houses. Each chassis is capable of holding up to 42 units (blades) in them. All the cases are stacked in a place called a server farm.

Each blade usually has the basic equipment of a computer, which are
    * memory to read input commands and data
    * a processor to perform commands manipulating that data, and
    * memory to store the results.

Other types of blades can be installed, they are used to provide switching, routing, and storage.

Advantages

Consider the advantages that blade servers have over conventional servers:

-Easy expansion: You can add or remove blades from the case easily without worrying about wires, and most of the blade servers are compatible with each other.

-Lower cost: Having blade servers remove the need to have individual servers with their wires and connections, they also save space and reduce the amount of energy used for cooling because the whole frame is used to cool the blades inside instead of cooling each server alone. 

-Easier management: Each case manages all the blades inside, so there is not any need to manage each server individually.  

-Easier maintenance: If any blade crashes, it does not affect the other blades and can easily be replaced by another one. Removing and installing a new blade is easy and can be done in a short time.

-Lower power consumption: New blade servers can save up to 75% on the energy used on servers now, some of it is on cooling and the other is the amount of electricity used. The chassis for blade servers uses much less electricity than the same amount of servers if they were put individually. Since the power supplies and other common components are not duplicated for each blade, this saves space and money. It can also increase reliability and decrease energy consumption.
 
-Better utilization of physical space: Instead of having many servers and connecting those using wires and a lot of space, blade servers connect in the case they are in and the cases connect together in the server farm. The server blade can offer double or triple the density of a standard rack-mount server. Each case does not need an individual monitor, so it also saves the space of monitors.


Blade server computing is fast transforming the server industry with a host of advantages in terms of design, functionality and total cost of ownership. By separating the CPU and memory from other components such as cabling, power supply, network connectivity and cooling systems, blade servers significantly reduce massive server farms into highly compact and dense blade servers. According to market research company IDC, the blade server market is projected to represent nearly 29 percent of server unit shipments worldwide by the end of 2008.

Applications

Blade servers' applications evolved from only used as a data center or for pack up to high-performance computing purposes such as distributed computing, rendering/imaging and data analysis, and web hosting. In addition, if newer blades or processors came out, you can change the old blades with them or have them working together. New blade servers have the ability to have more than one operating system, each blade can have its own operating system.

Differences between blade servers and supercomputers

While blade servers have a lot of CPUs working together, it can not match the number of processors in a supercomputer.  Each case has 42 processors on average. Blue Gene/L, the fastest supercomputer in 2007, has above 120000 processors in it. So if we have a blade with dual core processor, we need about