Knowledge Base

Q&A

1

100-Pair Cable Cutter

A cable and wire cutter built for larger diameter cabling. This cable cutter can be used to cut large diameter coaxial cabling such as RG9U, 100 pair group telephone cabling, and solid or stranded cables up to 2/o American Wire Gauge (AWG) wire.

1000Base-x

The physical layer standard for gigabit Ethernet connections used to connect switches to routers or other switches over fiber, and sometimes copper.

100Base-SX

A fiber version of gigabit Ethernet for short wavelengths (transmitting at 850nm)

100Base-T

An early UTP Ethernet standard that transmits data at 100Mbps.

100Base-TX

100 Mbps Fast Ethernet physical layer standard over UTP.

12U Server Rack

A smaller Server Rack with 12 Rack Units

19 Inch Rack

A 19 inch rack refers to a Server Rack dimension that was established by AT&T nearly 100 years ago and is defined by the Electronic Industries Alliance (EIA) to be compatible with almost all IT equipment. Its mounting fixture consists of two or four vertical posts which are each 0.625 inches wide and separated horizontally by 17.75 inches and a height of 1.75 inches, providing that the overall rack width (or “19 Inch Rack Shelf”) is 19 inches. 19 Inch Racks can vary in height, determined by the number of Rack Units it consists of, as well as in depth.

2

2 Post Rack

An Open Frame Rack, often referred to as a Network Rack or a Relay Rack often constructed of a lighter weight aluminum used mostly for vertically mounted network equipment and Patch Panels.  These are often used by Network Service Providers and can be found in Meet Me Rooms.

3

3rd Party Data Center

See:  Colocation Data Center

3rd Party Data Center Provider

See:  Colocation Provider

4

4 Post Rack

Most commonly referred to as a Server Rack, a 4 Post Rack is designed for deeper, heavier IT equipment than 2 Post Racks. It is constructed for horizontal mounting of IT equipment such as Servers and Blade Chassis. 4 Post Racks are typically constructed of steel and can exceed 52 Rack Units.

40U Server Rack

A Server Rack with 40 Rack Units.

42U Server Rack

The most common Server Rack height consisting of 42 Rack Units.

A

A-Side Customer

A term used in the colocation industry, and throughout UpStack to identify the Customer that requests and pays for interconnection services from a colocation company.  The A-Side Customer requests an interconnection to another customer, that Upstack refers to as the  Z-Side Customer.

Aggregate Power

The total power required by a building which is the sum of  IT power, essential power, and other common power loads. This is the total power required from the utility to a site, building, or data center.

Agile

A methodology used in software development that focuses on iterative and incremental improvements for software releases to test/find product market fit.

American Wire Gauge (AWG)

Also known as the Brown & Sharpe wire gauge, AWG is a wire conducter gauge standard in use since 1857 in North America. It is use to gauge the diameters of copper wire.

Amp

A steady state current that is created by applying one Volt across one Ohm of resistance.  To calculate amps in a data center, divide the Watts by Volts.  Watts / Volts = Amps.

Ansible

Ansible is an open-source platform written in Python for Configuring, Deploying, Automating, and Orchestrating IT workloads. Unlike Puppet or Chef it uses SSH, rather than an agent on the remote host.

Application Service Provider

A precursor to today’s more popular term, SaaS Provider, whereby applications are hosted, managed, and licensed by the provider.

Automatic Transfer Switch (ATS) PDU

For colocation customers that have hardware without redundant power supplies, rather than buy new equipment to achieve redundancy at the hardware level, or to eliminate downtime during PDU maintenance, you can use an ATS PDU that automatically switches to a secondary power source if the primary power source to the PDU is terminated. The dual AC power inputs can be connected to two power sources, whether they are utility, generator, or UPS. In most data centers, the primary power input plugs into the primary UPS system, which must be an On-Line UPS System. The secondary input plugs into the secondary UPS system. Ideally, the UPS systems should be connected to separate circuits. The PDU monitors voltage and frequency, automatically switching to the secondary input if the primary input deviates from the designated operating.

Automation

The process of scripting and programming software and tools to allocate compute, storage, network, and applications without human intervention. Automation enables IT workloads to be deployed and ready for use within minutes rather than hours, days, or weeks.

Availability

When evaluating IT, colocation, and network services, it is important to understand how well the system or services has historically performed.  Availability refers to the period of time that a data center, computing resource, or application is available for use and not offline for maintenance or a failed state. UpStack measures Availability as the percentage of time that the services are available over a period of time. Availability should not be confused with Reliability. Whereas Reliability represents the predictability of how a system is designed to perform, Availability is a measure of how a system has actually performed historically. Similar to an aircraft that is designed to be highly reliable, if the maintenance on that aircraft is inadequate, or the pilots responsible for operating the aircraft are not qualified, that highly reliable aircraft may crash.

B

Bandwidth

The amount of capacity you have to transmit data on a network – whether it is between two locations or multiple locations.

Bare Metal Servers

See:  Dedicated Servers

Base Rent

The minimum contracted payment owed by a customer. In NNN Colocation contracts, in addition to Base Rent, the customer is responsible for paying certain pro rata costs associated with operation of the data center and the building within which it operates; this would include all expenses such as property taxes, insurance premiums, repairs, utilities, maintenance, security, and direct power consumption.  In Modified Gross Colocation contracts, the Base Rent will typically include all such expenses, with the exception of the power consumed by the customer’s IT equipment and its pro rata share of power required to cool its IT equipment. In Gross Colocation contracts, the Base Rent is the absolute total payment to cover every expense related to the customer’s colocation requirement.

Basic PDU

A Basic PDU is a cost-effective way to deliver reliable power to IT hardware from a single power input (most often in data centers from a UPS system) with multiple output receptacles (breakers) that take higher voltage (often 480v) power from a UPS system and convert it to a lower voltage (often 208v or 120v) which is the voltage required by IT hardware. A basic PDU is the simplest way to distribute data center power from a UPS or non-UPS-supported power supplies to multiple Server Racks.

Biscuit Jack

Also known as a “Biscuit Block” jack, a Biscuit Jack is a device used for mounting network cabling with RJ45 connectors.

Blade Chassis

Sometimes referred to as a Blade Enclosure, a Blade Chassis is a device mounted into a Server Rack that can hold multiple Blade Servers.  Unlike Rack-Mounted Servers that are mounted directly to a Server Rack and require individual Power Supplies, Fans, and Networking harware, Blade Servers are installed in a  Blade Chassis that centralizes power, cooling, and network hardware across all of the Blade Servers installed within it.

Blade Server

A modular Server, that does not consist of the power, cooling, and network equipment found in traditional Rack-Mount Servers.  A Blade Server is installed within a Blade Chassis that centralizes power, cooling, and network equipment to support multiple Blade Servers installed within the Chassis. As a result of minimizing the amount of components needed for each server, and centralizing them on a common Blade Chassis, Blade Servers may minimize physical space requirements in Server Racks (and data centers) and increase the power utilization of each Server on the aggregate.

Branch Circuit

Often referred to as a “whip,” “power whip,” or “PDU cable,” a branch circuit is the final electrical circuit between the outlets in a Power Strip located on a Server Rack and the circuit breaker located in a RPP or PDU.

Branch Circuit Monitoring

Branch Circuit Monitoring enables colocation providers to meter/monitor multiple Branch Circuits to multiple colocation customers within a data center. This enables colocation providers to monitor the power consumption of its customers at the Server Rack for the purposes of tracking power circuit utilization and passing through electrical consumption costs to customers with contracts that provide for electrical pass throughs, such as Triple Net and Modified Gross contract structures.

C

Cable Raceway

See:  Ladder Rack

Cable Retaining Post

A Ladder Rack accessory that provides extra cable depth for additional cabling that mounts to an existing Ladder Rack.

Cable Runway

See:  Ladder Rack

Cage

See: Colocation Cage

Capital Expense (CAPEX)

CAPEX refers to funds used to build, acquire, or upgrade a physical asset such as servers, server racks, generators, or buildings. These expenses can include anything from purchasing or repairing a power cable to constructing an entire data center.

Carrier Dependent Colocation

A form of Data Center Colocation whereby the network service provider (or “Carrier”) provides colocation services within its own data center exclusively to customers of its network services.

Carrier Ethernet Exchange

Compared to older technologies, Ethernet provides lower cost per bit, greater bandwidth scalability, and simpler operations. A Carrier Ethernet Exchange enables Ethernet buyers and sellers to discover, transact and interconnect to each other, enabling service providers to cost-effectively scale Ethernet deployments and expand services for customers, and businesses to scale networks to keep up with application and bandwidth demands. Data centers that offer Carrier Ethernet Exchange services  give buyers and sellers of Ethernet services an easier and less costly way to connect and extend their Ethernet network reach.

Carrier Neutral Colocation

In contrast to Carrier Dependent Colocation, where a carrier itself is providing colocation services dependent on using its own network, a Carrier Neutral Colocation data center permits any carriers to become colocation customers within its data center to provide network and other services to other customers located within the data center. Carriers utilize a centralized Meet Me Room (MMR) where they interconnect with other colocation customers. The density of carriers within a singular data center often drives down network expenses for colocation customers resulted in a significantly lower TCO.

Cat5 Cable

Cat5 Cabling is a UTP type of cabling with an RJ-45 Connector used to connect computers and servers to modems, networks, and ISPs. Although the Cat5 cable can handle up to 10/100 Mbps at a 100MHz bandwidth (which was once considered quite efficient), Cat5 has become obsolete in recent years due to newer versions of Cat cables Cat5E and Cat6 that transfer data at much higher speeds with designs with significantly less Crosstalk.

Cat5E Cable

“Cat5 Enhanced” cabling replaced Cat5 as the standard cable about 15 years ago and is still the most commonly used in today’s data centers.    Cat5E can transfer data up to 10 times faster than Cat5 (gigabit Ethernet) at 100 MHz with a significantly greater ability to traverse distances without being impacted by crosstalk.

Cat6 Cable

Category 6 cabling arrived just a few years after Cat5E cabling and is ten times faster (10Gb at 250MHz). Further, it has an internal separator that further mitigates the impact of crosstalk. Cat6 cabling is most often used for network backbones due to its distance limitations and cost factors. If a Cat6 cable distance exceeds 164 linear feet, its speed rapidly decays to the same gigabit ethernet speed of Cat5E.

Cat6A Cable

Category 6A cables are the latest iteration of UTP ethernet cables that further mitigates crosstalk through exceptionally thick plastic material and extends its 10Gb speed beyond the limitations of Cat6 cables to 328 feet.

CCTV

See: Closed Circuit Television

CDN

See: Content Delivery Network

Ceiling Clear Height

Clear Height in Colocation Data Centers refers to the lowest portion of the ceiling in the data center that may interfere with racks with varying heights. It is important to confirm that the lowest portion of the ceiling in the proposed Colocation Data Center meets the height requirements for Server Racks.

Central Processing Unit (CPU)

Also known as the Processor, or the “Chip,” the CPU processes the information and instructions used by a computer. It is basically the brains of a computer.  The speed with which a CPU processes data and instructions is measured in either megahertz or gigahertz.

Children’s Internet Protection Act (CIPA)

CIPA is a federal law enacted to address concerns about access to offensive content over the internet by minors on library and school computers. It requires K-12 schools and libraries to incorporate filters and other measures to mitigate the risk of minors from receiving or retrieving harmful content

Children’s Online Privacy Protection Act (COPPA)

Instituted by the FTC in 2000, COPPA applies certain requirements on website operators to enforce privacy for children under 13 years old.  The primary goal is to put parents in control of the information that websites collect from their children.

Chip

See:  Central Processing Unit

Circuit Breaker

A Breaker is an electrical switch designed to protect an electrical circuit form damage caused by overload or short circuit. It automatically interrupts the power circuit in case of failures. Once the problem has been resolved, a Breaker can be reset to resume normal operation.  Breakers come in a wide variety of sizes and specifications, depending on the circuit that they need to protect.

In wholesale colocation, the customer is often required to purchase and install its own breakers since the colocation vendor will typically not know the individual circuits required to each server rack.

Closed Circuit Television (CCTV)

The most common type of video surveillance used in data centers. These are the video cameras you often see in many areas within and around the perimeter of data centers that transmit signals to monitors located in security offices within the data center. These cameras may be designed to pan, tilt, zoom, and can be constantly monitoring and saving footage or can be activated due to motion detected within its range.

Cloud Broker

A single integrated digital platform that aggregates multiple Anything as a Service (XaaS) cloud providers. A Cloud Broker platform enables executives within an organization to log into a single portal to broker and administer multiple XaaS offerings to one or more departments within the organization, helping to manage challenges with Cloud Sprawl. Standard features include a centralized cloud management portal, reporting, and system and data portability between downstream cloud providers.

Cloud Computing

Cloud Computing is a model for accessing computing resources (processors, memory, storage) over the Internet or via private network circuits in and between data centers and other facilities. There are five key characteristics of cloud services as defined by NIST which are a) on-demand self-service, b) broad network access, c) resource pooling, d) rapid elasticity, and e) measured service.  There are multiple cloud deployment models including Public Cloud, Private Cloud, Virtual Private Cloud, Community Cloud, and Hybrid Cloud.

Cloud Native Application

An application designed specifically to run on and benefit from cloud computing infrastructure.

Cloud Servers

Also known as virtual servers, Cloud Servers enable customers to utilize or rent virtual portions of a physical server. Instead of renting the full capacity of a physical, dedicated server, customers can rent smaller virtual portions of that same physical server. With Cloud Servers, customers only pay for on-demand increments of server capacity and have the benefit of infinite elasticity, scalability, and flexibility to scale IT resources up, down, in, or out depending on need and avoid paying for idle infrastructure or underutilized resources.

Cloud Sprawl

The uncontrolled proliferation of cloud services being consumed within an organization by its employees. When employees or departments purchase and consume various SaaS and IaaS products without proper controls in place. Examples include the marketing department using Adobe Cloud, engineering using AWS, accounting using Intuit Cloud Services, sales using Salesforce.com, and others using gmail, dropbox, etc. This can lead to compliance concerns, unnecessary license fees, and IT’s inability to gauge total organization IT workloads. Cloud Broker models have been created to control and better manage the cloud services that organizations consume.

Co lo

See:  Data Center Colocation

Co-location

See:  Data Center Colocation

Coaxial Cable (COAX)

Coaxial cable is used as a transmission line for radio frequency signals and provides much greater bandwidth than POTS.

Colo

See:  Data Center Colocation

Colocation

An arrangement whereby a business (customer) pays a 3rd party (Colocation Provider) for use of floor space or rack unit space in a data center for the customer’s IT equipment such as server racks and servers, an allocation of power for its IT equipment, and interconnection to other service providers, networks, or ISPs.  The Customer effectively outsources the complete operation of the physical building, security, and electrical and mechanical infrastructure to a 3rd Party that specializes in designing, building, and operating data centers. The Customer owns and operates everything that resides within its floor space, and the Colocation Provider is responsible for ensuring that its IT equipment never fails due to environmental events or power failure.

Colocation Cage

As the name suggests, a Colocation Cage is a physical barrier within a multi-tenant data center that adds additional security for customers by separating them from other customers in the colocation data center.  Although the electrical and mechanical systems may still be shared between other customers in the data center, some customers prefer to pay for this additional layer of security to mitigate the risks of other customers endangering their equipment.

Most colocation vendors will provide customizable cages that surround customer racks and limit access to only those approved by the customer. A typical Cage consists of cage walls, and one door with a standard mechanical lock. Customers can typically customize the security using additional layers of security leverage Biometrics, Retina Scans, Keycards, etc. There is typically a minimum footprint or minimum number of Rack Equivalents required for a dedicated cage to a customer.

Colocation Interconnection

In carrier Neutral Colocation facilities, customers are able to connect with networks, carriers, ISP’s and other customer by virtue of Interconnection. This enables customers to interconnect with global networks and ISPs for peering, transit and traffic exchange requirements, as well as other service providers that provide private cloud, public cloud, and financial services. This is often charged by the Colocation provider on a non-recurring and monthly recurring basis. For a deeper dive check out Understanding Colocation Economics.

Colocation Suites

A customer Colocation Suite is a dedicated room for Colocation Customers. It is physically separated from other Colocation Customers by floor to ceiling walls. Customers with high security requirements will often prefer a Colocation Suite to a Colocation Cage. Although a customer in a Colocation Suite will be physically separated from other customers, it does not mean that the infrastructure supporting the Colocation Cage is dedicated to the customer. It is important to determine what infrastructure (if any) is truly dedicated to the customer within a Colocation Suite.

Computer Room Air Conditioner (CRAC)

A CRAC unit is a common type of cooling system in data centers that.  They are the large units often seen on the perimeter of a data center room. A CRAC unit is similar to the cooling system in homes and consists of a Direct Expansion (D/X) refrigeration cycle that is powered by a compressor that is within the unit along with fans that draw air across a cooling coil pumped with refrigerant.  This type of cooling is one of the least energy efficient cooling designs in data centers today, but also the lowest upfront CAPEX for data center operators.

Computer Room Air Handler (CRAH)

A CRAH unit is a common type of heat rejection (cooling) system found in data centers. These units can be most often found on the perimeter of a data center room.  Not to be confused with a CRAC unit that utilizes a compressor and D/X cooling system within the unit, a CRAH unit has a cooling coil supplied with water (instead of refrigerant) from a chiller plant that is used to reject the heat from return air that enters the CRAH unit from the exhaust of servers and other IT equipment.

Configuration Management System (CMS)

Process for maintaining the configuration of your systems throughout their life cycle to enable a repeatable process required for automation and infrastructure management.

Content Delivery Network (CDN)

A CDN consists of proxy servers that are geographically distributed in multiple locations (data centers or POPs) to deliver highly available and low latency content to its customers.

Cooling Factor

Also known as a “Cooling Multiple,” and often in correlation with the PUE of a data center, the Cooling Factor is an additional cost added to Wholesale Colocation contracts to account for the power used by the cooling system to cool a colocation customer’s IT equipment.

When adding a Cooling Factor to a customer’s monthly bill, the colocation vendor will submeter a customer’s power consumption, typically via Branch Circuit Monitoring or at the PDUs serving the customer’s colocation environment, and will multiply the total kWh during the billing period by the Cooling Factor, to account for the additional cooling expense. For example, a persistent 100kW IT load over 30 days would be 72,000kWh (100kW x 720hours = 72,000). If the Cooling Factor is 1.5 (a 50% cooling factor) then the total power consumption passed through to the customer is 72,000 x 1.5 =  108,000kWh.  If the Normalized kWh Rate is $.06/kWh the total cost to the customer would be $6,480 (108,000 x .06).

COTS

Commercial-off-the-shelf software that is packaged and sold via multiple channels and made widely available from software companies.

CPU

See:  Central Processing Unit

CRAC Unit

See: Computer Room Air Conditioner

CRAH Unit

See:  Computer Room Air Handler

Cross Connect

A common term used to describe the type of interconnection provided between two colocation customers. There are many different types of cross connects such as Local Cross Connects, Metro Connect, Campus Connect, and Innerduct Connect.

Crosstalk

Crosstalk refers to the interference that occurs when data cables are in close proximity to one another, such as when they are in a cable tray or server rack. Crosstalk increases errors and lost packets (among other issues). Due to its propensity for Crosstalk, Cat5 cables have been replaced by newer versions of cat cables (i.e. Cat5E, Cat6, and Cat6A cables) that reduce the impact of crosstalk through a variety of methods, such as better shielding and UTP design.

D

Data Center Colocation

See: Colocation

Data Center REIT

A REIT specializing in owning and operating powered shell and/or colocation data centers.

Data Center Tiers

The Uptime Institute created and maintains a set of ratings applied to the design, construction, and operation of enterprise and colocation data centers. It defines a range of four distinct standards (or data center tiers). These Tiers are Tier 1 Data Center, Tier 2 Data Center, Tier 3 Data Center, and Tier 4 Data Center.

Data Centre

See:  Data Center

Data Sovereignty

The concept that computing data is subject to the laws of the country in which it is located. This prevents foreign governments form subpoenaing the host country or cloud service provider where the data center is located.  Data sovereignty presents technical challenges for companies that move data from on-premises to a cloud service provider. Privacy and data sovereignty laws vary by country. For example, the strictest countries of Germany, France, and Russia dictate that all of its citizens’ data is stored within the country’s physical borders.

Datacenter

See:  Data Center

Datacentre

See:  Data Center

Dedicated Server

Also referred to as Bare Metal Servers, a Dedicated Server is a physical server that is purchased or rented exclusively for the business. Unlike Cloud Servers, where physical servers are virtualized to provide shared, virtual environments on-demand for short-term, or unpredictable workloads, Dedicated Servers are utilized by businesses that require higher levels of data security, have compliance requirements, or have more predictable, long-term workloads.

Demarc

See:  Demarcation Point

Demarcation Point

In colocation, a demarcation point (also referred to as a “demarc”) defines the hand-off point between a colocation company and a customer. The demarcation point can be at the individual server rack level, a colocation cage, or colocation suite, depending on the size and type of the colocation requirement.

For network interconnection, the demarcation point is the defined location within a patch panel located in a colocation cage or server rack.

 

 

Derated Circuit Breakers

A Derated Circuit Breaker value is the safe maximum continuous power draw for a circuit. While its Rated Value (also known as its “Maximum Capacity”) is its overload protection tripping threshold, a Derated value is the guideline established by NEMA for its maximum safe sustained power draw, which it has determined to be no more than 80% of its rated value.  This provides a 20% safety margin into circuit usage during transient peaks as well as increased power usage during boot up compared to avg operational power usage.

Circuit breakers in data centers are temperature sensitive. They function on the thermodynamic properties of metals and if that breaker is exposed to hot or cold conditions it will not function properly.  As a result, electricians must follow NEMA guidelines to derate their circuit breakers depending on the application, installation and intended usage.

For rack-level power in data centers located in North America, the circuit breaker within the PDU for this circuit must be derated by 80% per NEMA ratings. So, this circuit, is in fact, not a 3.6kVA circuit. It is actually a 3.6kVA x 80% = 2.88kVA circuit.  To convert this to its kW equivalent, it depends on the efficiency of the power supplies ultimately delivering the power to the IT hardware.

In Europe and other parts of the world, circuits are simply described at their Derated capacity which helps to avoid confusion as to what capacity is truly available to a customer. It is important in the U.S. to ensure that you are quoted IT Power circuits at their Derated capacity.

DIACAP (DoD Information Assurance Certification and Accreditation Process)

DIACAP was a US Department of Defense (DoD) process to ensure that risk management is applied on information systems. Any service provider that stores, transmits, or display any DoD data must be DIACAP compliant. In 2014, the DoD transitioned the DIACAP process to the Risk Management Framework (RMF).

Direct Current (DC) Power

The majority of power circuits in the data center deliver Alternating Current (AC).  Direct Current (DC) is the native power used in all electronics (CPU, disk drives, memory chips). For equipment running directly on DC power, they do not need components to transform the power from AC to DC, which results in less failures and more efficient power distribution.

Dry-Pipe Sprinkler System

See:  Pre-Action Sprinkler System.

E

EC Plug Fan

A new, more energy efficient fan used to reject heat in data centers. EC stands for “Electrically Commutated” which in the most simple terms, is a type of motor driven by DC current instead of AC current that is able to manage the speed of the fan by regulating the current.

The term “plug” refers to the location of the fans in a cooling unit, unlike traditional CRAC/CRAH units where the fan is located inside of the unit, a plug fan is attached to the bottom below a raised floor.

Economizer

Mechanical apparatuses use to reduce the energy consumption of cooling systems such as air-side economizers and water-side economizers.

Elasticity (scale in/out)

The ability of an application or computing resource to automatically “scale in or out” by adding or removing virtual machines or computing resources to manage an increase or decrease in workload.

End of Row Switching

An alternative approach to network switching design whereby, rather than placing an Ethernet Switch in each Server Rack, as in the Top of Rack design, a specific rack is placed at the end of a row of Server Racks for the purpose of aggregating multiple switches within that rack to provide network connectivity to the servers within the row.

EPO

See:  Emergency Power Off.

Essential Power (Cooling) Load

The watts required by cooling equipment to reject the latent heat dissipated from IT equipment and other internal loads in a data center. Essential load is required for calculating aggregate power load and PUE.

Ethernet

The name coined by Xerox and defined by the IEEE 802.3 subcommittee for a group of network technologies and protocols commonly used for connecting computer systems to form LANs.

ETSI Rack

An ETSI Rack is a standardized Network Rack form factor dictated by The European Telecommunication Standards Institute (ETSI) aimed at defining the housing equipment used to accomodate optical fiber distribution and other netwoork equipment including Splicing and Patching Shelves and Distribution Sub-Racks.  An ETSI Rack consists of 4 width dimensions of 150 mm, 300 mm, 600 mm, or 900 mm and 2 depth dimensions of either 300 mm or 600 mm.

F

Failed State

A state for a system or a component of that system that suffered from a failure and cannot perform useful work.

FedRAMP (Federal Risk and Authorization Management Program)

A US federal standard approach to security assessment, authorization, and continuous monitoring of cloud products and services. FedRAMP is a set of cloud security standards to enable faster service procurement by all of the US agencies that adopt the approach and is also used by non governmental entities to quickly evaluate cloud service providers.

FISMA (Federal Information Security Management Act)

FISMA mandates that all US Federal Agencies are required to develop, document, and implement programs to provide information security for the systems that support operations and assets managed both internally and by other agencies or third parties.

Floor Live Load

The Floor Live Load capacity of a data center determines the amount of forces that cause stresses on a building or structural element, including equipment, people, vehicles, etc. and are measured as pounds per square foot.

The weight of a loaded Server Rack itself can weigh in excess of 2,000lbs. For this reason, when evaluating a data center it is important to understand whether the building’s design Live Load can support the customer’s IT equipment. For data centers with raised floor environments, it is important to not only know the subfloor live load rating, but also the total tile load of the raised floor tiles that will be supporting rack.

One of the most overlooked areas in data centers is Floor Live Load rating. Today’s IT systems are demanding higher ratings and some legacy data centers may not be able to support new form factors such as taller racks that can stack greater amounts of hardware within the same square footprint.

Full Service Contract

See:  Gross Contract

G

Gaseous Fire Suppression

A type of fire suppression that uses chemical agents and/or inert gases such as FM-200 to extinguish fires. This type of fire suppression is commonly found in data centers to mitigate the risk of damaging equipment with water to extinguish fire.

Gross Colocation Contract

Also known as a Full Service Contract, this contract type is commonly found in retail colocation requirements and consists of an all-inclusive rental rate that includes all operating expenses, power costs, etc. whereby the customer shares no risk for any increase in operating expenses over the base amount.  In this type of contract the customer knows exactly how much it will pay each month without the risk of additional costs due to increases (or decreases) in utility rates, taxes, operating expenses, etc. Because the colocation vendor is taking on the risk of additional increases in expenses, there is usually a risk premium included in the base rent which results in substantially higher base rental rates as compared to Modified Gross and NNN contracts where the risks are shared wholly or in part by the customer.

H

Half Rack

A Server Rack or Rack Level Requirement that is the physical height equivalent of one half of a Standard Rack Level Equivalent, or, roughly 22 Rack Units.

HIPAA (Health Insurance Portability and Accountability Act of 1996)

US legislation that provides privacy and security protection of sensitive patient information. All companies that manage patient information must incorporate security measures both physically and digitally for services they provide.  HIPAA compliant IT service providers, such as cloud and hosting providers strictly in place and followed.

Hot Swap PDU

A Hot-Swap PDU is able to make an On-Line UPS System hot-swappable. Much like a Basic PDU, The primary power input plugs into an on-line UPS system, and the secondary input plugs into a wall outlet. However, if the UPS system is taken offline for maintenance, repair or replacement, the PDU keeps the load powered by automatically switching from the primary input to the secondary input. When the UPS system is restored, the PDU will switch back to the primary input.

I

Idempotency

In contrast to an immutable  process whereby a resource is created or destroyed when an automated script is re-run,  Idempotency refers to the ability of a system to achieve the same results each time an automated script using the same resources, is re-run.  Configuration Management Solutions such as Ansible and Chef are examples of Idempotent technologies.  When automating scripts this helps removes fears of damaging active systems during automation.

Immutability

I contrast to Idempotency where resources are modified, Immutability is the concept whereby resources are never to be modified, resulting in new resources created and existing resources to be terminated.  Cloudinit and Terraform are two examples of Configuration Management Solutions that are Immutable.

Infrastructure Masons

A non-profit organization aimed at improving transparency in the data center industry through community involvement and industry workshops. The group has created the Data Center Performance Index.

Internet Exchange

An Internet Exchange facilitates Peering via an ethernet switching fabric. An Internet Exchange enables parties to peer to reduce network costs and enable efficient IP traffic interchange by aggregating thousands of networks directly at the exchange onto a shared fabric connecting peers at multiple data centers.  This enhances network performance by eliminating network hops and can reduce bandwidth costs.

ISP

See:  Internet Service Provider

IT Power

IT Power (also referred to as “Critical Power” or “Critical Load”) and measured in Watts, Kilowatts or Megawatts refers to the Real Power required (or consumed) by the IT equipment of a colocation customer. In Colocation agreements, a customer typically identifies the total kW required in each Server Rack and is charged a fee to reserve that power for the customer’s exclusive use. IT power is different from Cooling Power (also known as “Essential Power Load“).

K

kVA

Kilo Volt Ampere stands for 1,000 Volt Amperes.

kW

See: Kilowatt

kWh

See:  Kilowatt Hour

L

Ladder Rack

Also known as a Cable Raceway or Cable Runway,  a Ladder Rack is a structure shaped like a ladder that is used to distribute and organize network cables such as Fiber, Coax, Cat5, Cat6, etc. They are typically mounted to the ceiling of a data center directly above rows of Server Racks but can also be mounted to walls, floors, or Server Racks. Ladder Rack installation is simple and requires little trade experience. To avoid routing cable without damage, many accessories are included with Ladder Racks such as 90-degree bends, waterfalls and cable retaining posts.  A Ladder Rack is not intended for Branch Circuits or other electrical cabling due to the risk of electromagnetic interference with the network cables.

Local Cross Connect

A Local Cross Connect is a type of interconnection available within a single data center that connects a colocation customer with another company or service provider within the same data center via fiber, copper, coax, etc.

M

Meet Me Room (MMR)

A Meet Me room is a special room or cage in a data center where customers can physically connect to one ore more services providers to exchange data and can also be used to interconnect two cages within the same data center. The Meet Me Room typically consists of 2 Post Racks with patch panels, and network switching equipment.

Megahertz (MHz)

unit of frequency, equal to one million cycles per second.

Megawatt (MW)

One million (1,000,000) watts.

Metered PDU

A metered PDU locally monitors load levels to avoid potential overload conditions based on a digital current meter

Metered Power Billing

See:  Circuit-Based Pricing.

Metro Connect

A type of interconnection for Colocation customers that want to connect to a company that does not reside in its specific data center but is in another data center within the same metropolitan area, Colocation providers will often provide a highly reliable connection between the two data centers via owned or leased fiber facilities. This is often offered via Layer 1 and switch-based Layer 2 services.

Modified Gross Colocation Contract

A colocation contract whereby a customer is charged (i) Base Rent for colocation services (usually an allocation of floor space and power capacity), and (ii) pass through expenses for the power consumed by IT and cooling equipment. A modified gross colocation contract includes pro rata operating expenses as part of its Base Rent.

Monitored PDU

A monitored PDU is similar to a metered PDU, but goes an additional step by providing the customer to remotely monitor each electric circuit over an internet or dedicated network connection (usually via SNMP).

Monthly Recurring Cost (“MRC”)

MRC is the minimum monthly charge, also known as Base Rent, that a customer contractually pays a colocation, network, or other service provider.

MRC

See:  Monthly Recurring Cost.

MW

See:  Megawatt.

N

National Institute of Standards and Technology (NIST)

Native

UpStack uses the term Native to define a service that is physically available within a specific data center via a Local Cross Connect. A service or company that is native to a specific data center enables customers within that data center to directly connect to them within the same physical location without requiring a network connection outside of the physical data center.  Having a service availability native in a colocation data center can result in better performance, security, and cost to companies that utilize the service and are within the same data center.  For companies with Hybrid Cloud architectures, it is often beneficial to directly connect to cloud service providers within the same data center that supports your on-premise IT infrastructure.

Network Circuit Turn-Up

The process of testing and activating a newly installed network circuit. In colocation, after a cross connect is installed and the colocation customer has extended the circuit to its IT equipment, a Turn-Up is scheduled to verify that the network connection is properly installed and that the network circuit is functioning correctly.

NNN

See:  Triple Net.

Non Recurring Cost (“NRC”)

A one-time fee charged to colocation customers to cover the expenses related to the customer’s new or additional service. Examples of NRC’s include charges to cover the installation of server racks, ladder racks, cable trays, patch panels, cross connects, and branch circuits.

O

Open Frame Rack

A 2 Post Rack or 4 Post Rack widely used for intermediate distribution points, cabling and patching frames that is not fully enclosed such as a Server Cabinet. Open Frame Racks provide simple, low cost mounting for Rack Mount Equipment. For data center users where additional physical security at the rack level is not required, Open Frame racks are a cost effective solution.

P

PCI DSS (PCI Data Security Standard)

PCI DSS is a standard used to enhance the security of credit card data and to encourage the global adoption of consistent data security enforcement. PCI DSS applies to all entities involved in payment card processing – including merchants, processors, acquirers, issuers, and service providers, as well as all other entities that store, process or transmit cardholder data.

PDU

See: Power Distribution Unit

PDU Cable

See:  Branch Circuit.

Peering

In order to make the internet efficient, individual carriers, ISPs and network operators need to exchange data by granting others access to their network. This process is known as peering which can be on a payment basis or on a free basis. Peering is an alternative to the traditional method of purchasing IP transit to reach partners across the Internet. By enabling data to bypass intermediary networks, peering delivers optimized end-to-end network performance, speed and reliability, while lowering overall IP transit costs.

Plain Old Telephone Service (POTS)

POTS is a voice-grade service employing analog signal transmission over copper lines and is the basic form of residential and small business service connection to the telephone network.

POTS

See: Plain Old Telephone Service.

Power Factor Corrected (PFC) Power Supplies

Power Factor Corrected Power Supplies use a circuit design technique to increase the Power Factor of a piece of IT equipment so that it approaches or reaches Unity Power Factor.  To understand how much real power (kW) a power circuit is capable of delivering to IT equipment, you must understand what the Power Factor for the Power Supply in the equipment is.  The Power Factor defines the efficiency of the circuit.  When Power Supplies are 100% efficient, they are understood to be at Unity Power Factor.  However, when Power Supplies are less than 100% efficient, the potential power of the circuit (kVA) must be multiplied by the power factor to understand how much real power capacity (kW) can be delivered by the circuit.  For example, when Power Supplies are at Unity Power factor, kVA = kW.  However, when they are not, kVA will be less than kW because kV * Power Factor = kW.

Power Strip (Rack Mounted PDU)

Unlike a floor-mounted PDU, that enables colocation providers with an important power distribution bridge between a building’s primary power and the racks of its customers at much larger amounts of energy (250kV+), a rack-mounted PDU mounts directly to a customer equipment rack. Whereas a colocation provider operating the data center will be able to control, monitor, and balance power loads to multiple racks, rooms, or data centers, a rack-mountable PDU enables end-users to control, monitor, and balance power to specific servers, switches, and other hardware.

Power Usage Effectiveness (PUE)

PUE is a metric developed by The Green Grid to measure data center infrastructure efficiency.  In short, PUE = Total Facility Power / IT equipment Power.

Data centers primarily consume power for two specific purposes: 1) the IT equipment requires power to operate (known as “Critical Load“), and 2) the IT equipment generates heat that needs to be rejected with cooling systems that require power (known as “Essential Load” or “Mechanical Load”).  Simply put, the more energy efficient your electrical distribution and cooling systems are, the less power (and cost) is required by the entire data center.

PUE can be used to 1) determine the total amount of power required from the utility, 2) determine the size of primary and secondary electrical infrastructure to support total data center power requirements, and 3) estimate the total electricity costs that are often passed through to colocation customers in contracts.

It is very important when evaluating colocation data centers, to understand if any electrical costs are passed through. In Modified Gross or NNN contracts, it is critical to understand how the role of efficiency plays into your Total Cost of Ownership

 

Power Whip

See:  Branch Circuit.

Powered Shell

Also known as a Powered Base Building, a Powered Shell is a commercial building that has the appropriate zoning and entitlements already in place for data center use, and has aggregated an abundance of power and often an abundance of networks into the building or to the site. Powered Shell’s are not data centers themselves, but due to the amount of time it may take to approve a property for data center use and/or increase the power capacity to a site or building, a Powered Shell greatly improves the speed-to-market for companies that need to quickly develop a data center within the Powered Shell.

Processor

See: Central Processing Unit

PUE

See:  Power Usage Effectiveness

R

Rack and Stack Services

A service consisting of a 3rd party installing IT equipment on/within a customer server rack or network rack. The service provider typically requires a diagram, such as a Visio elevation that defines guidelines for the equipment installation.

Rack Level Equivalent

The square inches of Usable Data Center Floor Space required for different types of Server Rack form factors. This metric is used to gauge how many potential Server Racks of a specific dimension are able to fit within the Usable Data Center Floor Space of a Data Center taking into consideration Columns, aisles, ramps, and other miscellaneous power and cooling equipment that may be within the data center space. UpStack uses a 19 Inch Rack as its Standard Rack Level Equivalent.

Rack Level Requirement

A standard developed by UpStack that defines the total space and Critical Power required by one customer Server Rack. Due to variations in Server Rack form factors and power densities, the Rack Level Requirement (“RLR”) enables end users and vendors to clearly define the data center requirement on a common basis and avoid more ambiguous and generalized qualifiers such as Watts per square foot.

Rack Mount Power Strip

See:  Power Strip.

Rack Mount Power Supply

Rack Mount Server

A Rack Mount Server is a Server designed to be installed into a physical framework known as a Server Rack. A Rack Mount Server is installed into a Server Rack with multiple mounting slots, each known as a Server Rack Shelf, designed to accommodate the Servers and other IT equipment. It is a horizontal form factor that is able to conserve physical floor space within a data center.

Rack Mounted PDU

See:  Power Strips

Rack Power Density

The amount of IT power, expressed in watts or kW, required by the  IT equipment within a single server rack.  To calculate the rack power density you aggregate the IT power required by all IT equipment mounted in a server rack or network rack. A common rack power density is 4-6kW. Rack Power Density typically does not account for the essential power also required by the cooling systems that reject the heat dissipated by the IT power.

Rack Unit

A unit of measure within a Server Rack defined as a height of 1.75 inches (44.45 mm).  1U is defined as 1 Rack Unit, where as 2U is 2 Rack Units. Most Server Racks consist of 42 Rack Units.

Raised Floor

A raised floor is a common type of flooring in data centers. A raised floor is structurally elevated (often 24″ – 36″) above a solid substrate (typically a concrete slab). It can be designed to serve a number of purposes such as to distribute electrical branch circuits from breakers to server racks, plumbing for cooling equipment, and most often for distributing cold air used for rejecting the heat dissipated from the IT equipment in server racks.  In a hot aisle/ cold aisle design in a data center, server racks are laid out in rows where a common “cold aisle” that has perforations in the raised floor tile allows cold air to be directed into the aisle.  That air is then sucked into the intake of IT equipment via the fans within the equipment and rejects the heat dissipated from the IT equipment into a “hot aisle” on the other side of the rack which has tiles without perforations.

Rated Circuit Breaker

A Circuit Breaker’s rated value is equivalent to its maximum Current Rating. For example, a 30A breaker connected to a circuit will be rated at 30A and any draw on the circuit that exceeds 30A for whatever reason, at anytime, will trip the breaker and terminate the circuit to protect it from overloading. If you’ve ever seen a breaker, the number you see printed on it is its Rated Value.

Rated Value (for circuit breakers)

See:  Rated Circuit Breakers

Real Estate Investment Trust (REIT)

A company that owns income-producing real estate and often trades on public stock exchanges to provide investors with very liquid real estate investments.

REIT

See:  Real Estate Investment Trust.

Relay Rack

See: 2 Post Rack.

Reliability

Reliability is the probability of any components or systems to perform without failure in accordance with its specifications. IT workloads, services, and applications require varying levels of Reliability and when evaluating data centers and services provided within those data centers, it is important to be able to predict, based on its design, how well the facility or service is expected to perform. Whereas Availability is a measure of how well the data center or service has actually performed historically, Reliability is a predictor of how well the data center or service is supposed to perform.  When evaluating data centers and services it is important to evaluate both Reliability and Availability.

Remote Hands

Remote Hands assists customers that need remote access to their equipment for performing simple trouble-=shooting or maintenance tasks. Remote Hands involve the most basic activities of a colocation vendors’ staff, that are very high level. It is very important to understand the experience and service required from a colocation vendor as these outsourced services vary widely.

Examples of remote hands include: power cycling (turning power on and off); toggling a switch; pushing buttons; securing cabling to connections; observing, describing or reporting on indicators or display information on equipment or consoles

Remote hands do not include: any service requiring any hardware expertise; any service requiring any tool or diagnostic equipment; opening equipment; movement of connected equipment

Retail Colocation

Colocation services consisting of IT power requirements less than 250kW and/or space requirements less than 50 rack equivalents.

Risk Management Framework (RMF)

In 2014, the US Department of Defense consolidated its security standards such as DIACAP, to focus on maximizing reciprocity and reducing duplicative efforts within its departments. RMF is the unified information security framework for the entire federal government that is replacing the legacy Certification and Accreditation (C&A) processes within federal government departments and agencies, the Department of Defense (DOD) and the Intelligence Community (IC).

RJ45 Connector

A network cable connector type with eight-wire UTP connections that is often found in 10/100/1000BaseT networking and also referred to as an 8P8C connector.

S

Server Rack

A Server Rack is an open frame or enclosure (a “Server Cabinet”) used for mounting Rack Servers (or “Rack-Mount Servers,” “Rack Mount Chassis”) and other IT equipment.  Server Racks can come in many different form factors designed to accommodate various types of IT equipment, the most common of which is a 19 Inch Rack. Most 2 Post Racks and 4 Post Racks are designed to conform to the 19 Inch Rack standard.

Side Cutting Pliers

Pliers used for cutting copper wire or trimming leads that are full flush cut.

Simple Network Management Protocol (SNMP)

Often used in DCIM, SNMP is a protocol that collects rich data from various managed devices in a network.

Single Phase Power

Single Phase Power is a two wire Alternating Current (AC) power circuit. It is common in data center requirements with IT equipment that requires less power in each Server Rack.

Smart Hands

Smart hands is an hourly service provided to colocation customers consisting of on-site technicians for installation, remote management, and technical assistance of IT infrastructure within a  data center.

Some examples of smart hands include: management of equipment components (ie cards, drives memory), installing replacing or removing devices and server racks, rack and stack services, troubleshooting, cable management, power cycling, and carrier circuit testing.

SNMP

See:  Simple Network Management Protocol

SPECpower_ssj2008

SPECpower_ssj2008 is a benchmark product developed by the Standard Performance Evaluation Corporation (SPEC), a non-profit group of computer vendors, system integrators, universities, research organizations, publishers, and consultants. It is designed to provide a view of a sever system’s power consumption running Java server applications.Our calculations are based on the first industry-standard benchmark that evaluates the power and performance characteristics of volume server class computers. It is available from the Standard Performance Evaluation Corporation (SPEC) a (https://en.wikipedia.org/wiki/SPECpower)and was introducted in December, 2007.

For a full reference of assumptions please see our SPECpower summary table.

SSAE 16

Standard Rack Level Equivalent

The UpStack standard for measuring Sellable Racks as well as average Rack Power Density. A Standard Rack Level Equivalent consists of the square inches of Usable Data Center Floor Space required for a 42U, 19 inch wide, 36 inch deep Server Rack.

Suites

See: Colocation Suites

Switched PDU

With a Switched PDU, you are not only able to remotely monitor your circuits, but are able to remotely connect or disconnect your power circuits. This enables IT staff to avoid costly service calls or visits to the data center when hardware needs to be rebooted, custom power-on/off sequences, and load shedding of non-essential loads during power outages to extend battery backup runtime for critical loads. Unused PDU outlets can be electronically locked off to prevent the connection of unauthorized hardware.  With Switched PDUs the primary power input comes from the UPS and secondary power input is from a separate, non-UPS power source.

T

TDP

See: Thermal Design Power

Thermal Design Power (TDP)

The amount of power required by the IT hardware in your racks is more of an art than a science. Hardware vendors provide Thermal Design Power (TDP) values which estimates the Power Dissipation of the main components of a system such as the CPU Type, CPU Clock Speed, Hard Drives, Video Cards, etc. Power Dissipation is the process by which hardware consumes power (Watts), and dissipate this energy through switching devices within the hardware (such as transistors or vacuum tubes) and the energy lost in the for of heat due to the impedance of the electronic circuits.

TDP, however, is only a specification of the amount of heat dissipated from the components to gauge the amount of heat rejection (cooling) required for the aggregated components within the hardware.

For a more detailed analysis please see our simple Peak and Average Server Power Calculator.

Must determine the peak and average power (kW) required by the IT equipment in your racks (See: Rack Level Requirement)

Three Phase Power

Three Phase Power is a three wire Alternating Current (AC) circuit mostly used for higher power density Server Racks.  Unlike a Single Phase circuit, a Three Phase Power circuit provides 1.732 (the square root of 3) time more power than the same voltage using a single phase circuit.  For Example, a 20A, 120V Single Phase Power circuit is rated for 20A x 120V = 2,400VA whereas the same 20A, 120V Three Phase Power circuit would be 20A x 120V x 1.732 = 4,157VA.

Throughput

The actual transmission speed that data is transmitted from one user to another; throughput is usually lower than a network’s rated speed.

Tier 1 Data Center

A tier 1 data center is the most basic, and least resilient of the four standardized data center tiers developed and marketed by the Uptime Institute as a gauge of reliability for data center design and construction.  In order for a data center to be qualified and marketed as a tier 1 data center, it must be certified by the Uptime Institute.  Most colocation data centers are not certified by the Uptime Institue and therefore cannot be marketed to customers as being any of the Uptime Institute’s data center tiers. In short, a tier 1 data center is one which has a dedicated room for IT equipment and uses a UPS system and generator for secondary power.  Customers evaluating data centers may often see data centers marketed as “Tier 1 Equivalent” or “Tier 1 Compatible” but must critically evaluate these data centers to determine if the design truly exceeds or conforms to the Uptime Institute Tier Ratings.

Tier 2 Data Center

A tier 2 data center is one of the four standardized data center tiers developed and marketed by the Uptime Institute as a gauge of reliability for data center design and construction.  In order for a data center to be qualified and marketed as a tier 2 data center, it must be certified by the Uptime Institute.  Most colocation data centers are not certified by the Uptime Institue and therefore cannot be marketed to customers as being any of the Uptime Institute’s data center tiers.  A major gating factor to determine if a data center meets or exceeds the requirements for a tier 2 data center is if there is redundancy built into components such as UPS systems, generators, and PDUs. Customers evaluating data centers may often see data centers marketed as “Tier 2 Equivalent” or “Tier 2 Compatible” but must critically evaluate these data centers to determine if the design truly exceeds or conforms to the Uptime Institute Tier Ratings.

Tier 3 Data Center

A tier 3 data center is the most common of the four standardized data center tiers developed and marketed by the Uptime Institute as a gauge of reliability for data center design and construction.  In order for a data center to be qualified and marketed as a tier 3 data center, it must be certified by the Uptime Institute.  Most colocation data centers are not certified by the Uptime Institue and therefore cannot be marketed to customers as being any of the Uptime Institute’s data center tiers.  A major gating factor to determine if a data center meets or exceeds the requirements for a tier 3 data center is if it is Concurrently Maintainable – that each and every component of its electrical and mechanical infrastructure has resiliency (and/or redundancy) built into the the components to enable any portion of the system to be taken offline for maintenance purposes without affecting the power or cooling of the data center. Customers evaluating data centers may often see data centers marketed as “Tier 3 Equivalent” or “Tier 3 Compatible” but must critically evaluate these data centers to determine if the design truly exceeds or conforms to the Uptime Institute Tier Ratings.

 

 

Tier 4 Data Center

A tier 4 data center is the most reliable standard of the four standardized data center tiers developed and marketed by the Uptime Institute as a gauge of reliability for data center design and construction.  In order for a data center to be qualified and marketed as a tier 4 data center, it must be certified by the Uptime Institute.  Most colocation data centers are not certified by the Uptime Institue and therefore cannot be marketed to customers as being any of the Uptime Institute’s data center tiers.  A major gating factor to determine if a data center meets or exceeds the requirements for a tier 4 data center is Fault Tolerance, Active-Active, and System Plus System for its electrical and mechanical distribution paths- that each and every component of its electrical and mechanical infrastructure. Customers evaluating data centers may often see data centers marketed as “Tier 4 Equivalent” or “Tier 4 Compatible” but must critically evaluate these data centers to determine if the design truly exceeds or conforms to the Uptime Institute Tier Ratings.

Top Of Rack Switching

A methodology of network switching whereby the Servers within the rack connect to one or more Ethernet switches mounted on the Server Rack. The term “top of rack” does not necessarily denote the actual physical location of the switch itself. The Ethernet switch links the rack to the network with fiber connecting from the Server Rack to a common aggregation point.

Transport Layer

Layer 4 in both OSI and TCP/IP that governs communication between two hosts.

Triple Net (NNN) Colocation Contract

A contract type whereby a colocation customer is charged a base rent for services provider by the colocation provider (often a reservation of floor space and power capacity) that passes through the costs of the customers power consumption for IT equipment and cooling equipment, as well as operating expenses such as common utilities, cleaning, water, security, maintenance are passed through to the customer.

Turn-Up

See:  Network Circuit Turn-Up

U

UL Listing

Underwriters Laboratory (UL) is the standard for safety and compliance for innovative electrical and mechanical products.

Uninterruptible Power Supply (UPS)

Unity Power Factor

Unshielded Twisted Pair (UTP)

A type of wiring in which  four twisted pairs of wiring are embedded within a sheath inside a network cord.

Uptime Institute

The Uptime Institute is a subsidiary of The 451 Group that provides consulting, certification, and training for stakeholders in the colocation data center industry. The Uptime Institute has created proprietary methods for standardizing data center designs at the physical layer that enables colocation customers to validate the reliability of a data center’s design and construction (See: Uptime Institute Tier Ratings), as well as its operations and maintenance procedures (See: Uptime M&O). The most well known Uptime Institute standard is its data center Tier Certification program that certifies which tier a data center is designed and constructed as.

UTP

See: Unshielded Twisted Pair.

V

VA

Volt Amperes

Virtual LAN (VLAN)

An abstracted LAN which consists of a closed collection of clients and servers isolated to the same VLAN.  The broadcast signals are partioned and isolated from other signals on the same LAN at the data link layer (OSI layer 2).

Virtualization

The process of segmenting a physical machine (the “host”) to run multiple virtual versions (see: instances) of something which, from a user experience, performs as if it were an actual physical machine.

VLAN

See:  Virtual LAN

Voice over IP (VoIP)

The transmission of voice signals over an IP Network.

VoIP

See:  Voice over IP.

VPC

See: Virtual Private Cloud

VPN

See: Virtual Private Network

W

WAN

See: Wide Area Network

Water-Side Economization

A type of economizer, enabled by wet and dry bulb temperatures that are low enough to allow a cooling system to reject heat without the need of a compressor or chiller by utilizing a wet cooling tower or a dry cooler.

Waterfall

A Ladder Rack accessory shaped like a waterfall that guides cabling from the ladder rack in a downward direction in a “waterfall” like fashion.

Watts Per Square Foot

The average IT power density per usable square foot in a data center.  This number does not include watts for essential (cooling) power. For example, if a data center is 20,000 gross square feet with 10,000 usable data center square feet, and the UPS capacity is 1,500kW, then watts per square foot would be on average 150 across the entire data center floor (10,000 usable data center square feet / 1,500,000 watts = 150 watts per square foot.)Watts per square foot is commonly used in the data center industry, but is not as accurate as understanding the watts per rack equivalent needed that can fit in the usable floor space.

Wet Pipe Sprinkler System

A common fire suppression system where the sprinkler pipes are pre-filled with water. Once the sprinkler heads are opened due to heat from a fire, the water is immediately discharged through the sprinklers.  Wet Pipe systems are not common in data centers and should be avoided and replaced with dry-pipe systems along with pre-action or gaseous fire suppression.

Whip

See:  Branch Circuit.

Wholesale Colocation

A term often used in the 3rd party data center sector to describe larger, and longer-term colocation requirements.  Unlike Retail Colocation, used to describe smaller colocation requirements that are typically structured as Master Service Agreements with Gross Pricing Structures, Wholesale Colocation typically consists of contracts in the form of real estate Leases with varying pass through expenses such as Critical IT Power Pass Through Expenses, Cooling Power Pass Through Expenses, and other Pro Rata Operating Expenses. Although the actual size may vary, Wholesale Colocation contracts typically begin at a minimum of 250 kilowatts of Critical IT Power.

Wide Area Network

A network that links different sites, such as data centers and corporate offices, to transmit data over long distances.

Z

Z-Side Customer

In colocation interconnection, the Z-Side Customer is the target customer that another colocation customer, referred to as the A-Side Customer, is requesting a connection to via a cross connect of some type.