A blade server, sometimes referred to as a high-density server, is a compact device containing a computer used to manage and distribute data in a collection of computers and systems, called a network. Its role is to act as a conduit between computers, programs, applications and systems.
In general, a blade server consists of a chassis, or box-like structure, housing multiple thin, modular electronic circuit boards, known as server blades. They are called blades because of their ultra-thin shape. Each blade contains a single server, often dedicated to a single application. The information within blade servers is stored on a memory card or other memory device. In addition, the individual blades contain processors, memory, integrated network controllers, an optional Fibre Channel host bus adaptor (HBA) and other input/output (IO) ports. These are used to connect server blades to other server blade units within the system, or to connect individual blades to power sources.
Benefits of blade servers
Because most blade servers are dedicated, meaning they are focused on a single use application, professionals have more control over how they are accessed and used, and how data is transferred between devices. Some of the benefits of blade servers compared to their traditional predecessors include:
- Cooling: Each blade is cooled individually by fans. Additionally, because of their stackability, the servers can be kept in smaller air-controlled areas that keep all of the mechanical parts at a proper temperature.
- Management supervisor included: Unlike their predecessors, blade servers can be controlled and managed in tandem with other server units within a data center or network. The administrator of a stack of blade servers is able to balance the workloads spread over many individual server devices.
- Seamless movement within rack and minimal wiring: Organizations using blade servers can experience a reduction in cabling for blade server housing compared to larger models like box servers. Because the design is modular and compact, individual units can easily be transferred within or between systems. With less cables and parts to deal with, IT administrators and other professionals can spend less time managing the infrastructure of their data center and more time ensuring high availability, which is the maximization of how well the servers perform even after failure.
- Low power consumption: Servers within a rack are able to share a single power source, leading to a reduction in storage and power consumption costs compared to other server types. In addition, the compact nature of blade servers causes them to need very little power. This decrease in power consumption allows teams to maximize the efficiency of their space.
- Storage consolidation: Each blade in a blade server typically comes with one or two local ATA or SCSI For additional storage, blade servers can connect to a storage pool facilitated by a network-attached storage (NAS), Fiber Channel, or iSCSI storage-area network (SAN). The advantage of blade servers comes not only from the consolidation benefits of housing several servers in a single chassis, but also from the consolidation of associated resources (for instance, storage and networking equipment) into smaller architecture that can be managed through a single interface.
- Compact size: Unlike traditional rack servers, blade computing does not have minimum size restrictions. Because blade servers are made up of the smallest number of components required to a device to be considered a computer, they are able to fit in small, tight spaces and run well with other simplified servers as part of a larger overall structure.
- High-trust compatibility: The nature of servers that carry out a highly individualized task makes it possible for an organization to dedicate a single server entirely to mission-critical applications, or programs without which the entire company or project could not exist.
Uses of blade servers
To maximize their efficiency, blade servers are often dedicated to single-task functionality. Some example of tasks a server may perform include:
This article is part of
Server hardware guide to architecture, products and management
- File sharing: Any transfer of data between digital points or devices.
- Web page serving and caching: The processes of delivering web pages to visitors and temporarily storing the information on the website on the visitor's computer so it can be quickly pulled up and recalled in order to reduce wait time and stalling.
- SSL encryption of web communication: Ensuring that information that travels over an internet connection is secure from outside parties, viruses and attackers.
- Transcoding: Converting the code of web page content to move seamlessly between differently-sized and shaped devices or for other conversion purposes.
- Streaming: Transmitting audio and video content without interruption to allow viewing and listening in real time.
- Load balancing: Like most clustering applications, blade servers can be used to include load balancing and failover
- Virtualization: Blade servers can be used to create abstract versions of applications or real-life activities for digital use.
- Storage: The sleeker, more compact design allows a larger amount of information to be stored to support a larger number of applications working in unison.
Difference between blade servers and rack-mounted servers
Blade servers and rack servers are two common features of data centers and while very similar, they have some significant differences. Typically, rack servers, also known as rack-mounted servers, are created to fit vertically within a shelf-like structure called a rack, typically housed in a data center or server room. They are ideal for projects with low CPU requirements or small amounts of physical space.
The number of servers stacked within a rack-mounted system is completely variable based on the size and extent of an organization's project needs. Rack servers are fully equipped to work as a stand-alone computer and are able to run multiple, complex applications and processes, using powerful memory and CPU resources. Their modularity and configurability make them convenient to assemble and store. Without the need for a large chassis to hold the servers, the rack-mounted servers are ideal for projects that require a small number of servers between, for instance, two and ten.
Blade server technology is very similar to rack-mounted server technology. Blade servers, much like rack-mounted servers, are usually stored in a vertical stack in a data center or server room, though they tend to appear in much larger data centers for projects with higher power needs. One key difference is that the individual servers are highly stripped-down, with simple, sometimes singular functionality. This allows for multiple small, flat, horizontal blades containing individual servers to be stored within a chassis and stacked vertically. Unlike rack servers, blade servers generally do not function as stand-alone servers--they work best when placed with others in a chassis and combined into a blade service stack.
Another major difference is that each individual server is typically dedicated only to a single application, as it generally only contains CPUs, controllers and memory. Blade server systems are much easier to fix than larger rack-mounted servers because the modular parts can be repaired individually without disrupting the larger system--in other words they can be hot-swapped. Companies need to be proactive researching the company that makes the blade server model they purchase for the organization. When a company chooses a specific model of blade servers, they need to prepare for the possibilities of vendor lock-in, so switching between products is costly as it involves a complete overhaul of all the available equipment. Blade servers also require fewer wires and cables than typical stack-mounted servers, and are powerful processors compared to the space they occupy.
History of blade servers
The first servers were offered in the 1990s and soon after there was a need for dedicated servers, especially in industries with privacy and data safety requirements, such as healthcare. Over time they became more widely used in other industries, especially as security threats and viruses continued to evolve. The need for smaller, more compact and faster servers increased with each new company and technology.
Blade servers were invented in the 2000s by Christopher Hipp and David Kirkeby at their company, RLX Technologies in Houston. Both Christopher Hipp and David Kirkeby were former employees of Compaq, one of the early companies to sell computers and computer products. Hipp was an entrepreneur and inventor who was especially interested in the advancements of graphic design as technology improved. Kirkeby was his co-worker who helped to invent the patent for the blade server.
They created the first commercial blade server in 2001 to meet an industry need for more compact and efficient data storage technology. Eventually the blade server design was acquired and manufactured by Hewlett Packard and other companies.
In its original format, all parts of the blades could be hot plugged and fit into a modular design. Servers with a similar design were also sometimes referred to as space-saving servers.
Types of blade servers
Blade servers are categorized by type based on various traits, such as differing CPUs, amount of RAM and cache memory or connections.
Blade servers can be equipped with different central processing units such as Intel, Advanced Micro Devices (AMD), Motorola and Sun Microsystems. The CPU does the processing and computing for the entire system, and the more powerful the central processing unit, the more functionality the computer system is able to perform at a given time.
A blade server can also contain a wide variety of memory systems. The first is static RAM (SRAM), one of the earlier forms of storing information common in digital cameras and printers that stores data exactly as it is. Another is dynamic RAM (DRAM), a form of storage that updates as it works and is common in more complex systems like video game consoles. Lastly is double data rate synchronous dynamic random access memory (DDR SDRAM), most commonly used in servers and computers with a great deal of visual content. RAM is used to manage and store information and complete processing and calculations in real time.
While blade servers have a modular and uniform design that makes them easy to move between data centers, they can have a variety of outer connectors. For example, they may be equipped with an Ethernet output, a token ring, a fiber channel which can be up to 6 miles long or a fieldbus network protocol.
Additionally, blade servers may be connected to different storage connection methods, including the following:
- FireWire: FireWire is a type of port often used to connect digital cameras, external hard disk drives, phones, tablets and devices that require the use of high transfer rates. A FireWire connection can support up to 480 Mbps connection speeds.
- SATA: A SATA connection, or Serial AT Attachment connection, is a bus-style port that connects host bus adapters to hard drives, solid-state drives or other larger storage devices.
- SCSI: A small computer system interface connection is used to connect computer parts in a SCSI system and create ports of communication between them.
- DAS: Directed attached storage is within a computer itself, or directly connected to it, and is not part of an external storage network system. A common example is the internal hard drive of a computer or laptop.
- FC: An FC connection, or fiber channel storage area network (FC SAN) storage connection that allows network storage devices to share information with servers at high speeds with high performance.
- iSCSI: An iSCSI connection, or internet small computer system interface, is a storage connection based on IP and includes a storage system that is used to link various network storage locations within the same server system.
Parts of a blade server technology system
The basic parts of a blade server system include:
- A chassis- The outer case of one or more blade servers used to store all of the mechanical parts. Usually small and compact in shape.
- Blades- Small mechanical parts that hold servers and contain the memory systems to store data and serve information to the computer systems in a network. Sometimes shaped like super thin drawers stacked on top of other blades.
- Servers- The servers within the individual blades in blade server technology are housed within the chassis and contain all of the memory, processors and programs to carry out the individualized function they are designed to complete.
- Racks- Commonly used to store many individual blade servers in a physical location like a server room.
- A backplane- The area of a sever rack that connects all of the individual modules and servers via circuitry, including sockets, plugs, wires, switches and circuit boards.
Future of blade servers
While blade servers are a highly common aspect of many business server rooms, there are other technologies that may replace their commonality. Brick servers, which unlike a blade server do not require a chassis outer container, have many of the benefits of compactness and compatibility, while taking up even less space. Similarly, cartridge servers are small like printer cartridges and commonly used in various industries. In addition, modular hybrid servers address some issues with hard-wired connections, temperature management and ease of movement within a server system. As power consumption and thermal design power (TDP) increase with ever-improving technologies, the demand for servers that can be stored seamlessly and easily cooled continues to grow.