Data centers are complex facilities that house computer systems, servers, storage systems, networking equipment, and other IT infrastructure. They require extensive infrastructure and facilities to ensure the equipment operates reliably and efficiently. Some of the major facilities included in a typical data center are:
Power Systems
Power is one of the most critical components of any data center. Servers, storage and network devices require clean, stable electrical power supply to operate 24/7. Data centers have extensive power infrastructure including:
- Utility power feeds – Data centers are typically fed by multiple independent utility power feeds to provide redundancy. The power will come into the building at medium voltage (typically 12-25kV).
- Backup generators – Diesel or natural gas generators provide backup power in case of utility power outages. There are often multiple redundant generators.
- Automatic transfer switches – Transfer power automatically to backup generators if utility power is lost.
- Uninterruptible Power Supplies (UPS) – Provide battery backup power to bridge the gap between loss of utility power and startup of generators.
- Power distribution units (PDUs) – Manage and monitor power distribution to server racks and cabinets in the data center.
Having robust power systems prevents power outages and enables uninterrupted operation of the critical IT systems.
Cooling Systems
Data center equipment generates a lot of heat. The servers, storage and network gear need to be cooled continuously to prevent overheating and malfunction. Data centers use extensive cooling systems including:
- Air handlers – Supply cool air into the data center room.
- Computer room air conditioning (CRAC) units – The CRAC units cool the air supplied by the air handlers using chilled water or direct expansion cooling.
- Chillers – Chill water for use in CRAC units for data center cooling.
- Cooling towers – Release heat from chiller water.
- Raised floors – Allow cool air to be delivered through perforated tiles to server cabinets.
- Hot aisles and cold aisles – Isolate hot exhaust air from equipment and cold intake air to equipment to optimize cooling.
- Rack cooling devices – Spot coolers, fan trays etc. to provide supplemental cooling to racks.
Careful monitoring and control of data center temperature and humidity is critical to maintain optimal hardware operating conditions.
Fire Suppression
Fires can catastrophically damage a data center and cause extensive downtime. Data centers are equipped with highly sensitive smoke detection systems and fire suppression systems to extinguish fires before they spread. Fire protection includes:
- Very early smoke detection apparatus (VESDA) – Sample air continuously and detect trace amounts of smoke.
- Fire suppression gas – Typically FM-200 or Novec 1230 gas is used to rapidly suppress fire without damaging electronics.
- Pre-action sprinkler systems – Water sprinklers activated only if smoke is detected to prevent accidental water discharge.
Monitoring Systems
Data centers use monitoring systems to track numerous environmental parameters and alert staff if conditions become abnormal. Monitoring systems help prevent outages and failures by enabling early detection of issues. Some monitored parameters include:
- Temperature and humidity
- Electrical power status
- UPS and battery health
- Cooling system metrics
- Network connectivity
- Server and storage health
- Fire and leak detection
- CCTV surveillance
Monitoring systems can visualize data center conditions in real-time and trend data over time to identify anomalies.
Security Systems
Data centers store highly sensitive and critical data and infrastructure. Comprehensive physical security measures prevent unauthorized access and protect against threats. Security systems include:
- Access control system – Manage and log access to facility using card readers, biometric systems etc.
- Video surveillance – CCTV cameras monitor entry points, rooms and aisles.
- Intrusion detection systems – Motion detectors, door prop alarms and glass-break sensors detect unauthorized entry.
- Mantraps – Prevent tailgating by verifying access before entering facility.
- Bollards – Concrete posts protect facility against vehicular access.
Personnel access to data center floors is typically very restricted and limited to critical staff only.
Reliability and Redundancy
Achieving extremely high reliability is paramount for data center design. Downtime translates directly into lost revenue and productivity. Key concepts for reliability include:
- Concurrently Maintainable – All critical facility systems are redundant to enable maintenance without affecting operation.
- Fault-tolerant – Infrastructure continues operating seamlessly during component failures.
- No single point of failure – There is redundancy for all critical systems.
Some examples of redundancy include:
- Multiple utility feeds from independent substations
- Redundant generators, ATS switches, chillers, pumps etc.
- Fire suppression system redundancy
- Diverse physical pathways for fiber entry
Having redundancy requires additional upfront investment but helps avoid the high cost of downtime over the facility lifetime.
Structured Cabling System
Data centers have an extensive structured cabling system to interconnect equipment. This includes:
- Fiber optic backbone cabling
- Redundant telecom rooms on each floor
- Vertical cable runs to connect floors
- Horizontal copper and fiber cabling to racks
- High-density fiber panels
- Patch panels for network connectivity
A well-designed cabling system with adequate capacity enables flexibility in configuring network architecture as needed.
Mechanical Infrastructure
Data centers require extensive mechanical systems infrastructure including:
- Large electrical rooms with UPS systems and transformers
- Multiple large chilled water plants with pumps and chillers
- Air handling units and extensive ductwork
- Plumbing systems for cooling towers and sanitation
- Rigging/mounting provisions for heavy equipment
Space must be allocated in facility design for adequate mechanical rooms and provisions for necessary equipment.
Operations and NOC
Data centers have dedicated central control rooms for monitoring and managing the facility. This includes:
- Network operations center (NOC) – Coordinate technical operations of critical systems.
- Building management systems (BMS) – Centralized monitoring and control of mechanical and electrical systems.
- Work benches, conference areas and offices for staff.
The NOC and BMS enable staff to efficiently monitor and operate the data center from a central location.
Loading Dock & Receiving
Space is needed inside the data center to stage equipment deliveries before they are deployed. This includes:
- Large loading dock and bay doors for trucks
- Material handling equipment – Forklifts, pallet jacks etc.
- Staging and inventorying space
Proper loading facilities enable efficient receiving and deployment of racks, servers and other infrastructure.
Storage and Workshop Space
Support rooms provide space for inventory storage, maintenance workshops and equipment staging. These spaces include:
- Parts storage rooms and racks
- Battery storage rooms
- Tool cribs and maintenance workbenches
- Equipment repair and staging areas
Adequate support rooms enable technicians to efficiently maintain, repair and stage equipment.
Meeting Rooms
Meeting and collaboration spaces enable staff to work effectively. Data centers typically have:
- Conference rooms – For team meetings and conference calls.
- Huddle spaces – Smaller rooms for focused work.
- Breakroom – Staff lunch and breaks.
Providing onsite meeting rooms enhances productivity and minimizes security concerns of having external visitors.
Office Space
Data centers have office space for facility engineers, IT technicians, operations staff, security personnel and other employees. Typical spaces include:
- Private offices
- Cubicles and workstations
- Bullpen areas
- Technical libraries
- Lockers and restrooms
Onsite office space allows critical staff to be located proximate to the data center floor.
Conclusion
Data centers are complex Mission Critical Facilities with extensive infrastructure to provide highly reliable and secure operations. Key facility systems include power, cooling, fire protection, monitoring, security and cabling. Supporting facility spaces are also required including operations centers, loading docks, storage and office space. Careful planning and design is needed to incorporate all necessary data center facilities.