Servers are powerful computers that deliver applications, services and data to end-user devices. Data center servers come in several form factors:
Rack-mount servers are wide, flat standalone servers—the size of a small pizza box— designed to be stacked on top of each other in a rack, to save space (vs. a tower or desktop server). Each rack-mount server has its own power supply, cooling fans, network switches, and ports, along with the usual processor, memory, and storage.
Blade servers are designed to save even more space. Each blade contains processors, network controllers, memory and sometime storage; they’re made to fit into a chassis that holds multiple blades and contains the power supply, network management and other resources for all the blades in the chassis.
Mainframes are high-performance computers with multiple processors that that can do the work of an entire room of rack-mount or blade servers. The first virtualizable computers, mainframes can process billions of calculations and transactions in real time.
The choice of form factor depends on many factors including available space in the data center, the workloads being run on the servers, the available power, and cost.
Most servers include some local storage capability—called direct-attached storage (DAS)—to enable the most frequently used data (hot data) to remain close the CPU.
Two other data center storage configurations include network-attached storage (NAS), and a storage area network (SAN).
NAS provides data storage and data access to multiple servers over a standard Ethernet connection. The NAS device is usually a dedicated server with multiple storage media—hard disk drives (HDDs) and/or solid state drives (SSDs).
Like NAS, a SAN enables shared storage, but a SAN uses a separate network for the data and consists of a more complex mix of multiple storage servers, application servers, and storage management software.
A single data center may use all three storage configurations—DAS, NAS, and SAN—as well as file storage, block storage and object storage types.
The data center network, consisting of various types of switches, routers and fiber optics, carries network traffic across the servers (called east/west traffic), and to/from the servers to the clients (called north/south traffic).
As noted above, a data center’s network services are typically virtualized. This enables the creation of software-defined overlay networks, built on top of the network’s physical infrastructure, to accommodate specific security controls or service level agreements (SLAs).
Power supply and cable Management
Data centers need to be always-on, at every level. Most servers feature dual power supplies. Battery-powered uninterruptible power supplies (UPS) protect against power surges and brief power outages. Powerful generators can kick in if a more severe power outage occurs.
With thousands of servers connected by various cables, cable management is an important data center design concern. If cables are too near to each other, they can cause cross-talk, which can negatively impact data transfer rates and signal transmission. Also, if too many cables are packed together, they can generate excessive heat. Data center construction and expansion must consider building codes and industry standards to ensure cabling is efficient and safe.
Redundancy and disaster recovery
Data center downtime is costly to data center providers and to their customers, and data center operators and architects go to great lengths to increase resiliency of their systems. These measures include everything from redundant arrays of independent disks (RAIDs) to protect against data loss or corruption in the case of storage media failures, to backup data center cooling infrastructure that keeps servers running at optimal temperatures, even if the primary cooling system fails.
Many large data center providers have data centers located in geographically distinct regions, so that if a natural disaster or political disruption occurs in one region, operations can be failed over to a different region for uninterrupted services.
The Uptime Institute (link resides outside ibm.com) uses a four-tier system to rate the redundancy and resiliency of data centers:
Tier I—Provides basic redundancy capacity components, such as uninterruptible power supply (UPS) and 24/7 cooling, to support IT operations for an office setting or beyond.
Tier II—Adds additional redundant power and cooling subsystems, such as generators and energy storage devices, for improved safety against disruptions.
Tier III—Adds redundant components as a key differentiator from other data centers. Tier III facilities require no shutdowns when equipment needs maintenance or replacement.
Tier IV—Adds fault tolerance by implementing several independent, physically isolated redundant capacity components, so that when a piece of equipment fails there is no impact to IT operations.
Data centers must be designed and equipped to control environmental factors—most of which are interrelated—that can damage or destroy hardware and lead to expensive or catastrophic downtime.
Temperature: Most data centers employ some combination of air cooling and liquid cooling to keep servers and other hardware operating in the proper temperature ranges. Air cooling is basically air conditioning—specifically, computer room air conditioning (CRAC) targeted at the entire server room, or at specific rows or racks of servers. Liquid cooling technologies pump liquid directly to processors, or in some cases immerse servers in coolant. Data center providers are turning increasingly to liquid cooling for greater energy efficiency and sustainability—it requires less electricity and less water than air cooling.
Humidity: High humidity can cause equipment to rust; low humidity can increase the risk static electricity surges (see below). Humidity control equipment includes the aforementioned CRAC systems, proper ventilation, and humidity sensors.
Static electricity: As little as 25 volts of static discharge to damage equipment or corrupt data. Data center facilities are outfitted with equipment to monitors static electricity and discharge it safely.
Fire: For obvious reasons, data centers have must be equipped with fire-prevention equipment, and it must be tested regularly.