How do blade servers work?

Blade servers offer several advantages including a compact and modular design and small footprint because of their high server density. The ultra-thin server blades are housed in compact enclosures, ensuring high server density for optimum capacity, flexible IT administration and lower energy consumption.

Blade servers were invented in the 2000s by RLX Technologies in Houston. The first commercial blade server was developed in 2001 by Christopher Hipp, the ‘father of blade technology’, and David Kirkeby. As compact, scalable servers with modular design, they met the demands for space-saving, powerful servers for dedicated tasks.

Blade servers are considered one of the most compact server technologies available due to the server density, reduced cabling and efficient power consumption. This makes them suitable for optimal processing and distribution of workloads, consolidation of network and storage devices, and dedicated use for specific, business-critical applications. Blade servers offer flexible scaling of computing capacities due to their high, modular server density which also makes them suitable for use in spatially limited or large data centres with high energy demands.

The core of a blade server are the chassis and the blades. The chassis is usually an elongated, box-like enclosure that contains the thin blades. Blades are essentially printed circuit boards functioning as servers. They usually contain corresponding storage systems, processors, network controllers, input/output ports and hot bus adapters.

The structure of a blade server is as follows:

• Chassis: Box-shaped, small enclosure containing all electronic components including server and racks
• Blades: Carrier units containing servers and server components
• Server: Housed in individual, modular blades and comprising the servers including components such as processors, dedicated programs and input/output ports
• Racks: Serve as external racks for blade servers to be grouped and housed as a single unit in a server room
• Backplane: A component of racks through which the modular blade servers can be connected via cabling, power sources or printed circuit boards

Blade servers are used for a diverse range of space- and energy-saving functions. Since blade servers can be reserved for dedicated tasks due to their modular design, they are often used for the distribution of workloads or business-critical applications.

Common applications include:

• Virtualisation: Thanks to virtualization of apps and hardware, physical IT infrastructures can be reduced. Because of the flexible and compact design, blade servers are suitable for software and hardware abstraction to use IT resources more efficiently.
• File sharing: Blade servers perform classic server tasks such as backing up, restoring, and transferring files between endpoints and digital nodes.
• Cluster-computing und load balancing: Blade servers often serve as components in cluster computing and server clustering to provide high availability, scalability, load balancing and failover services.
• Web hosting: The modular server units provide flexible scalability and high availability for web hosting.
• File sharing: For dedicated tasks such as file sharing services, blade servers provide high availability and reduced downtime as part of computer clusters.
• Webpage caching: Blade servers can significantly reduce wait times and downtime when caching websites.
• SSL encryption and malware protection: Dedicated blade servers ensure a secure, encrypted data exchange and protect against malware.
• Streaming: Streaming audio and video requires uninterrupted data exchange and blade servers help to optimise real-time streaming.
• Storage: With more servers distributed across less space, blade servers offer plenty of storage while consuming less power.
• Website transcoding: For a trouble-free display of website content across different end devices, blade servers are suitable for code conversion of website elements and mobile SEO.

Blade servers differ depending on the equipment and manufacturer. Market leaders include well-known names such as Cisco Systems, Supermicro, Dell and HPE. The most important distinguishing features in terms of equipment include:

• Performance of CPU: Processors and processing units from manufacturers such as Intel, Sun Microsystems, Advanced Micro Devices and Motorola are often found in blade servers. These in turn determine how powerful the server units are.
• Storage media and working memory: For real-time information processing, blade servers depend on sufficient memory in the form of SSD or HDD and good random access memory (RAM). Blade servers combine different RAM systems, such as static RAM, which stores data in its original state, or dynamic RAM, which allows for on-the-fly updates. For servers that rely on processing visual data, DDR SDRAM is usually used as dynamic random access memory with doubled data rate.
• Connection options: Blade server connections can be made via a token ring, an Ethernet output, via fibre optic channels or a fieldbus network protocol.
• Connection options to storage systems: Blade servers can be connected to storage using different port types, including FireWire, SATA, SCSI, DAT, FC and iSCSI.

The alternatives to blade servers tend to be even more compact. They include brick servers, for example, which do not require a case at all and therefore take up even less space. Other server alternatives are cartridge servers that are similar in size to a printer cartridge. These are closely linked to micro servers usually intended for specific purposes and contain modular server cartridges.

However, blade servers and their alternatives are similar in that they strive to achieve a high and powerful server density via cluster computing while saving as much space and energy as possible. Basically, the server technologies of the future are becoming ever smaller and more efficient. Modular models offer more advantages in terms of maintenance, servicing and scaling compared to static server systems.

In addition to blade servers, rack servers are often used. They are in similar in that they feature a space-saving design for high computing power but differ in some essential aspects. Rack servers are usually mounted vertically in structures that resemble shelves. Like blade servers, the number of stacked rack servers can be expanded or reduced. Here, the server performance adapts to the CPU requirements, available space and individual projects. In other words, they are modularly designed and configurable systems, but their design is more static than that of blade servers.

Blade servers tend to be used in projects with high power requirements and dedicated applications. That’s where the differences between blade servers and rack servers become obvious. Blade servers usually consist of servers that are reserved for specific applications. The server blades in the chassis are usually stacked horizontally and used for one dedicated task per blade. In addition, unlike rack servers, individual blade servers usually do not act as main servers but are active as combined blade server stacks.

A major pro of blade servers over the somewhat larger rack servers is their ease of maintenance and repair. Server blades are easy to replace and maintained on a modular basis. The con is that companies can be dependent on the components of a particular manufacturer and thus suffer from a vendor lock-in. Nevertheless, with blade servers you save on complex cabling and benefit from a space-saving, flexible and efficient server technology.