Modern data centers and other computing facilities can have thousands of servers, input/output modules, routers, switches, and other types of processing units supported by a common utility infrastructure. For example, the utility infrastructure can provide power distribution that supply power to the individual processing units from a power grid, a battery bank, a diesel generator, or other power sources. In another example, the utility infrastructure can also include transformers, rectifiers, voltage regulators, circuit breakers, or other types of electrical/mechanical components that condition, monitor, and/or regulate the supplied power.
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.
In certain computing facilities, individual component enclosures can house multiple servers, input/output modules, routers, switches, and/or other types of processing units. Each component enclosure can also include a power distribution unit (“PDU”) that provides power to the processing units. The PDU typically can include one or more plugs, breakers, cords, receptacles, metal housings, and/or other electrical/mechanical components arranged in circuits that distribute power from a main power source to the individual processing units.
Certain components of PDUs may be location or power-system specific. For example, power systems in Europe are primarily 400-volt AC phase-to-neutral while those in the United States can be 208-volt AC phase-to-phase. As a result, PDUs and associated component assemblies for Europe may require different plugs, breakers, or other components than those for the United States even though the component assemblies may contain the same configuration of processing units. Such difference in PDU components may increase manufacturing complexities of component assemblies and/or component enclosures, and may also lead to production or deployment delays or installation errors.
Several embodiments of the present technology are directed to modular power distribution in which component assemblies of a single configuration of processing units may be suitable for multiple locations or power systems. For example, a PDU according to one embodiment of the present technology may be divided into a first subsystem that is location specific and a second subsystem that is assembly specific. The first subsystem can include plugs, breakers, cords, and/or other electrical/mechanical components arranged in circuits based on and suitable for particular voltage/current ratings, source topology, grounding requirements, and/or other power system characteristics. The second subsystem can include receptacles, cords, or other components that are independent of the characteristics of the particular power system, but are suitable to the particular requirements of the processing units.
The first and second subsystems can each include a connector configured to mate with each other. Each connector can include multiple conductors configured to allow electrical power to flow from the first subsystem, via the second subsystem, to the processing units. In certain embodiments, the connectors may be universal or common for all types of power systems. As a result, component assemblies with a single configuration of processing units may be manufactured for multiple locations irrespective of the particular power system characteristics at such locations. Thus, manufacturing complexities and production delays of component assemblies of processing units may be reduced when compared to conventional techniques.
Certain embodiments of the present technology can also reduce capital costs for upgrading data centers or other types of computing facilities. Unlike conventional techniques in which PDUs are fully replaced with component assemblies of processing units, certain components of PDUs in accordance with the present technology may be retained and reutilized during upgrades. For example, the first subsystem of the PDUs may be reused while only the second subsystem is replaced with component assemblies to accommodate upgraded processing units. As a result, the first subsystem may be depreciated over a longer period of time than the second subsystem, and thus reducing capital costs of facility upgrades.
Certain embodiments of systems, devices, components, modules, routines, and processes for modular power distribution in computing facilities are described below. In the following description, specific details of components are included to provide a thorough understanding of certain embodiments of the present technology. A person skilled in the relevant art will also understand that the technology may have additional embodiments. The technology may also be practiced without several of the details of the embodiments described below with reference to
As used herein, the term “power distribution unit” or “PDU” generally refers to an apparatus with multiple power outlets configured to supply and/or distribute electrical power from a power source to multiple electrical or electronic devices. PDUs may be floor mounted, enclosure mounted, rack mounted, or may have other suitable structural profiles. Certain example PDUs may contain one or more power conversion and/or conditioning components that condition and/or transform one or more larger capacity power feeds into multiple lower-capacity power feeds. Example power conversion and/or conditioning components include transformers, circuit breakers, power filters, and power rectifiers. In other examples, PDUs may simply include a number of appliance or interconnection couplers.
Also used herein, the term “processing unit” generally refers to an electrical or electronic apparatus configured to perform logic comparisons, arithmetic calculations, electronic communications, electronic input/output, and/or other suitable functions when supplied with electrical power. Example processing units can include computing systems (e.g., servers, computers, etc.), computing devices (e.g., logic processors, network routers, network switches, network interface cards, data storage devices, etc.), or other suitable types of electronic apparatus. Multiple processing units may be organized into a component assembly and be carried by a rack, rail, or other suitable types of support component. Also used herein, the term “rack” or “rail” generally refers to a frame or enclosure into or onto which one or more processing units may be mounted.
Also used herein, the term “connector” or “electrical connector” generally refers to an electro-mechanical device or assembly configured as an interface for coupling electrical circuits. A connector may include a housing that may have any of many mechanical forms. For example, a connector may include a plug or a socket that mates with the plug. In another example, a connector can be a coaxial connector, a Molex connector, or of other suitable types of connector. A connector may also include multiple conductors (e.g., wires) configured to carry power and/or signals. The conductors may be electrically parallel to and insulated from one another.
As discussed above, power source characteristics at different locations may require different component assemblies with different PDUs in accordance with conventional techniques. As a result, manufacturing of the component assemblies may be complex and prone to production or deployment delays. Several embodiments of the present technology divide a PDU into a location-specific subsystem and an assembly-specific subsystem. The location-specific subsystem may be configured to receive power from a power source with certain characteristics (e.g., 3-phase AC at 208 volts phase-to-phase) at a particular location and output a power supply of a particular configuration (e.g., single-phase AC at 120 volts phase-to-neutral). The assembly-specific subsystem may be configured to receive power from the power supply of the location-specific subsystem and provide the received power to the processing units. As a result, the assembly-specific subsystem of the PDU can be independent from the characteristics of the power source at the location, rendering the component assembly suitable for multiple locations with different power supply characteristics.
In the illustrated embodiment, the power source 130 includes a utility power grid with one or more characteristics. For example, the utility power grid may have at least one of a particular voltage rating, current rating, source topology (e.g., delta or wye), grounding requirement, plug specification, and/or other electrical/mechanical characteristics based on a location (e.g., country or territory), primary use (e.g., commercial or residential), and/or other suitable factors. As discussed in more detail below, a portion of the PDU 120 may be configured based on and/or corresponding to the one or more characteristics of the power source 130. As a result, the PDU 120 may receive and distribute power from the power source 130 to the processing units 104. Even though the power source 130 is shown in
The component enclosure 101 can have a size and dimension configured to contain the processing units 104. For example, though not shown in
The processing units 104 can be configured to implement one or more computing applications, network communications, input/output capabilities, and/or other suitable functionalities, for example, as requested by the users 101. In certain embodiments, the processing units 104 can include web servers, application servers, database servers, and/or other suitable computing components. In other embodiments, the processing units can include routers, network switches, analog/digital input/output modules, modems, and/or other suitable electronic components.
As shown in
The first subsystem 120a can be configured in accordance to the one or more characteristics of the power source 130 in order to receive power therefrom (herein referred to as “location specific”). In certain embodiments, the first subsystem 120a may include electrical components arranged in a circuit corresponding to the one or more characteristics of the power source 130. For example, the power source 130 may have a voltage rating of 400-volt AC phase-to-neutral in a delta topology. Based on such characteristics, the first subsystem 120a may include one or more plugs, breakers, transformers, or cords with ratings corresponding to 400-volt AC. The first subsystem 120a can also include suitable circuits that distribute the received power in a delta topology to multiple electrically parallel branches. In another example, if the power source 130 has a voltage rating of 208-volt AC in a wye topology, then components of the first subsystem 120a may have different ratings and/or specifications as well as different circuits to distribute the received power. One example first subsystem 120a is described in more detail below with reference to
The second subsystem 120b can be configured independently from the one or more characteristics of the power source 130 but instead based on characteristics of the processing units 104 (herein referred to as “assembly specific”) in the component assemblies 102. For example, the second subsystem 120b can include receptacles, cords, or other components arranged in circuits that correspond to configurations of the processing units 104 in the component assemblies 102. The receptacles, cords, or other components, however, can be selected, designed, and/or otherwise provided irrespective of the characteristics of the power source 130. Thus, the second subsystem 120b may be common or “universal” for most or all locations irrespective of the characteristics of the power source 130. As a result, component assemblies 102 with a single configuration of the processing units 104 may be manufactured for multiple locations, and thus reducing manufacturing complexities and production delays when compared to conventional techniques. One example second subsystem 120b is described in more detail below with reference to
The set of connectors 110 can be configured independently from the characteristics of the power source 130 to electrically connect the first subsystem 120a to the second subsystem 120b. The set of connectors 110 can mate with each other in any suitable fashion. For example, the set of connectors 110 can include a plug and a socket configured to mate with the plug. In the illustrated embodiment, the set of connectors 110 are shown as a first connector 110a associated with the first subsystem 120a and a second connector 110b associated with the second subsystem 120b of the PDU 120 located inside the component enclosure 101. In other embodiments, the set of connectors 110 may include multiple subsets of connectors or may have other suitable arrangements located outside the component enclosure 101 or at other suitable locations. One example set of connectors 110 is described in more detail below with reference to
In operation, the first subsystem 120a of the PDU 120 receives power from the power source 130. The first subsystem 120a can then distribute the received power into multiple branches. The set of connectors 110 can allow the distributed power to flow along the multiple branches from the first subsystem 120a, via the second subsystem, and to the processing units 104 in the individual component assemblies 102 of the component enclosure 101.
Certain embodiments of the computing framework 100 can reduce capital costs for hardware upgrades. Unlike conventional systems in which PDUs are fully replaced with component assemblies 102 of processing units 104, the location specific first subsystem 120a of the PDU 120 may be retained and reutilized during upgrades. Only the assembly specific second subsystem 120b may require replacement to accommodate upgraded processing units 104 in the component assemblies 102. As a result, the first subsystem 120a may be depreciated over a longer period of time (e.g., a 15-year depreciation cycle) than the second subsystem 120b (e.g., a 3-year depreciation cycle), and thus reducing capital costs during hardware upgrades. Certain embodiments of the computing framework 100 can also reduce inventory of hardware components. Conventional techniques may require an inventory of different component assemblies with different power configurations even though the component assemblies all contain the same configuration of processing units. In contrast, the present technology can provide component assemblies with a single configuration suitable for most or all locations with different power systems.
As shown in
As shown in
As shown in
Even though the receptacles 108 are shown in
Specific embodiments of the technology have been described above for purposes of illustration. However, various modifications may be made without deviating from the foregoing disclosure. In addition, many of the elements of one embodiment may be combined with other embodiments in addition to or in lieu of the elements of the other embodiments. Accordingly, the technology is not limited except as by the appended claims.