US20130191658A1 - Methods of achieving cognizant power management - Google Patents
Methods of achieving cognizant power management Download PDFInfo
- Publication number
- US20130191658A1 US20130191658A1 US13/790,601 US201313790601A US2013191658A1 US 20130191658 A1 US20130191658 A1 US 20130191658A1 US 201313790601 A US201313790601 A US 201313790601A US 2013191658 A1 US2013191658 A1 US 2013191658A1
- Authority
- US
- United States
- Prior art keywords
- power
- state
- server
- power utilization
- management
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3206—Monitoring of events, devices or parameters that trigger a change in power modality
- G06F1/3209—Monitoring remote activity, e.g. over telephone lines or network connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/324—Power saving characterised by the action undertaken by lowering clock frequency
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/3296—Power saving characterised by the action undertaken by lowering the supply or operating voltage
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5094—Allocation of resources, e.g. of the central processing unit [CPU] where the allocation takes into account power or heat criteria
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Definitions
- the invention relates to power management.
- a system and method is presented for optimizing power management of computational and electrical appliances provisioned in data centers and server installations.
- FIG. 1 illustrates major categories of energy consumption
- FIG. 2 is an exemplary architecture for power management in a server farm or data center environment according to the present invention.
- FIG. 3 is an exemplary framework related to the architecture shown in FIG. 2 .
- FIG. 4 is an exemplary architecture for power management in a virtual machine environment according to the present invention.
- FIG. 5 is an exemplary top level flowchart of a method according to the present invention.
- FIG. 1 depicts the major sections of energy consumption in server installations 100 .
- Total energy 110 is a summation of the power used for computational purposes 120 and the power used for heating and cooling functions 130 .
- Environmentally based power management uses these major consumption areas to minimize the total energy usage in an adaptive manner.
- the iterative analysis begins by computing the initial measure of total electrical power consumed by the server and network hardware configuration from the asset information.
- the assets denote the electrical and computing units in the infrastructure and cover, for example, servers, routers, switches, cooling units, racks, sensors and power distribution panels. This base computation is then supplemented by the electrical power delivered to the cooling and other infrastructure elements.
- This electrical side power consumption is tracked against the computational power consumed by business services and correlated with real time electrical metering at the server level.
- the asset information and other configuration information can either be supplied directly to the power management system or imported from existing IT management systems, which is common in an enterprise level installation.
- the environmentally conscious power management system and method of the present invention increases the efficiency of overall power utilization by using coordinated power monitoring at the electrical power distribution side as well as in the consumption side of server installations using simplified common management interfaces (e.g. simple network management protocol (SNMP) web services) for environmental and server monitoring.
- SNMP simple network management protocol
- the system and method improves the efficiency of power utilization per computational unit and provides policies for dynamic load allocation capabilities based on application loads and environmental contexts, e.g. rack configurations, cooling units and thermal characteristics of the data center room.
- the dynamic computational load balancing policies of the present invention can be adopted by (i) positioning the servers at environmentally optimized locations, (ii) scheduling the computational tasks at physically diverse locations, and (iii) providing VMOTION (utility for migration of virtual machines between physical servers offered by VMWARE) in the virtual environments leveraging virtualization technology.
- VMOTION utility for migration of virtual machines between physical servers offered by VMWARE
- the proposed approach addresses modifications needed to measure the electrical power requirements in such virtualization environments.
- a coordinated framework is the most important ingredient to maximize the benefits of information gathered by the environmental and server measuring agents denoted as monitors in the figures illustrated herewith.
- Practices address local electrical optimization which might not accurately determine the total requirements.
- multiple business services are consolidated at the server level, they might help to reduce the power consumed at each individual server.
- they do not account for the operating conditions and thus become ineffective in situations where those conditions are significant, such as when the servers are configured in a dense arrangement like server blades.
- an application running on a server with a rack mounted cooling unit would likely use less overall power than the same application running on an identical server in a less efficient cooling environment.
- an application running on a server with a lowpower hard drive would likely use less power than the same application running on a server with large power consumption or inefficient heat dissipation design.
- the selection of servers for load balancing is mostly ad hoc under current practices and does not detailed analysis.
- the optimization of overall power utilization thus does require a coordinated framework, systematic procedure and control elements to distribute the business computing load both at the physical location and logical locations (as applications and server clusters or virtual machines) by configurable and adaptive monitoring and continuous analysis of the environment for global power management decisions.
- the server monitoring tools and appliances employ either agent or non-agent based technologies to measure application load, starting with determining the number of instances of each application running on each server, using different mechanisms primarily depending on the operating systems hosted in server hardware.
- Typical monitoring interfaces available for monitoring and measurement include Windows management instrumentation (WMI) on Microsoft platforms, simple network management protocol (SNMP), and Web service management (WS-MAN).
- WMI Windows management instrumentation
- SNMP simple network management protocol
- WS-MAN Web service management
- the granularity of the information necessary to associate the application loads of the electrical power used is not directly computable from the standard interfaces. Thus it is useful to obtain the right metrics from these existing base metrics that are usable in electrical power computations.
- the environmentally-based power management framework addresses this constraint by aggregating the process level information from different mechanisms (including network computer monitoring devices such as Raritan's COMMAND CENTER NOC, baseboard management controller such as Raritan's KIRA, and power environmental monitoring devices such as Raritan DOMINION PX) to improve the accuracy of the predicted electrical power unit consumption based on environmental factors, server hardware characteristics, operating system overheads and running application to provide business services.
- network computer monitoring devices such as Raritan's COMMAND CENTER NOC
- baseboard management controller such as Raritan's KIRA
- power environmental monitoring devices such as Raritan DOMINION PX
- a first level of power metrics is obtained from server hardware and power supply units installed within the servers and a second level of power metrics is obtained from operating system and applications executed within the servers.
- a third level of metrics is computed from the live monitoring of actual business utilization, configuration, topology, thermal activity and electrical power usage.
- IPMI Intelligent platform management interface
- KIRA Raritan Baseboard Management Controller
- DOMINION PX DOMINION PX
- the environmentally cognizant power management framework also applies to both virtual and physical server environments.
- virtualization technology provides dynamic load balancing operations that help simplify collaborative power management.
- the proposed approach can be adopted to a physical server installations as well as virtual server installations.
- the management controller and power optimization ( 330 ) system of FIG. 3 supports the necessary components to order and group the data set collected over multiple interfaces to various time intervals.
- the information required to compute the long-term trending is different from localized short bursts of measurements.
- the management controller configures the necessary data collection devices to collect the measurements at different sampling intervals between maintenance cycles as necessary.
- System 200 includes servers 205 and 210 that are connected to a management and monitoring device 215 .
- Servers 205 and 210 may be any computing device in any hardware and software configuration that may include for example, server applications, database applications, web applications and the operating system that hosted business applications.
- Management and monitoring device 215 may be Raritan's Command Center NOC or any other similar device.
- Servers 205 and 210 are also connected to power management devices 220 and 225 , respectively.
- Power management devices 220 and 225 may be Raritan's DOMINION PX or any other intelligent power management device.
- Management and monitoring device 215 and power management devices 225 are further coupled to a monitoring control and analysis device 230 that is accessible to the user by a client interface 235 .
- the environmentally cognizant power management system and method uses the above management components as building blocks for power management framework.
- the management and monitoring device 215 determines in real-time the operating system and applications running a server.
- the management and monitoring devices 215 determines and real-time the operating system and applications monitoring on each server.
- the management and monitoring devices 215 monitor server hardware parameters using a Baseboard Management Controller's hardware that may be available as embedded modules or parts, such as for example, in Raritan's KIRA device. These parameters, including the power cycle events and CPU temperatures, provide an additional level of power metrics to correlate to the CPU utilization and the computational load created by the business applications.
- Power management devices 220 and 225 gather information of electric power used by each connected device as well as environmental data (primarily temperature, air flow and humidity) by wired and wireless sensor devices.
- This environmental monitoring integration grows as the energy requirements at the cooling units to the computing load and ambient conditions.
- the computational affects electric power draw which creates thermal energy which in turn affects the electrical power consumed by the cooling supply.
- the ambient conditions within the data centers affect the efficiency of cooling units, thus necessitating the need for including the localization parameters into the dynamic model of the data center.
- the external ambient conditions also impact the energy needs to operate cooling and condenser units at the desired operating temperatures and are preferably incorporated into the model as well.
- the business applications are distributed according to operating profiles computed by the centralized processing engine (main controller) running in control and analysis device 230 . That is, business applications are scheduled to be executed at a particular server based on its efficiency and power and thermal profiles and its position within the rack, location within the server room or facility, as well as environmental conditions.
- the processing logic in device 230 gathers environmental conditions from the environmental sensors which are coupled to a common monitor (or integrated into power management devices 220 and 225 ) and the measurement of electric power usage that are directly available from power management devices 220 and 225 .
- the framework takes into account the following:
- the measurement data are synchronized to a clock server that runs inside device 230 .
- the sequencing operation of the measurement data acquired from any device is validated based on the time of arrival of the relative event occurrences before storage for further processing.
- the Web services capability automates the discovery and communication between the information provider services and the processing engine which runs in device 230 .
- all measurement devices must be discovered and properly registered in order to successfully interpret the events with its context.
- the context of the operating conditions is important to the measurement. For example, the electrical power usage of computer room air conditioner (CRAG) units in a lightly loaded environment is to be treated differently from the measurement of highly loaded operating conditions.
- CRAG computer room air conditioner
- the data received throughout the entire system is thus interpreted within the operational context of the global level.
- the monitor data are now accessible to receivers, which, in the usual scenario, are provided in the main controller in the system. If there is a failure to receive data from one device or multiple devices, the management controller adapts the next level metrics or derived metrics from previous context or earlier data acquisition until the system is back to normal operation.
- the standard interface of using Web services fit well into scalable reconfigurable enterprise platform and the proposed approach can be easily adopted into its existing infrastructure.
- the prior art power management for information technology services does not provide the capabilities of coordinated operations over the network and thus cannot benefit from load-balancing across infrastructure and computational equipment in power utilization.
- the inventive system and method address this constraint by provisioning network aware services in each appliance.
- a network enabled Web services technology supports both reactive and proactive power monitoring, which is valuable to complex environments where the behavior of applications and short term failures or irreparable crashes.
- Web services can provide configuration updates and enable reconfiguration by subscription methods between peer-to-peer systems, which in the present case are the data measurement devices and data receivers for processing the data.
- server level power monitoring is highly beneficial for information technology management, specifically for auditing, capacity planning, thermal assessment, and performance improvement. These scenarios essentially require power usage and active monitoring on servers to determine the need for rescheduling the load based on predefined and/or dynamic power policies.
- the dynamic power management policies are either stored in the system or referred from an external network management system.
- Basic level of power profiles are supported at the operating system level and advanced configuration and power interface (ACPI) provides a way to control this aspect.
- ACPI advanced configuration and power interface
- the proposed approach applies to any power profiles including ACPI and uses appropriate metrics for different kinds of profiles.
- Current power policies that utilized ACPI focus on reliable power feed to the computing units based on the availability of electrical power from alternate sources like uninterruptible power supply generators.
- system 300 includes a power management appliance 305 that acts as a power distribution unit (PDU) to target servers 310 and enables users to measure power at the outlet level, providing an accurate view of power consumption at a target server 310 or a PDU level (total power distributed).
- PDU power distribution unit
- Most of the existing power distribution units measure rack level or branch circuit level power measurements, which might not be sufficient for power monitoring at the server level.
- management and monitoring device 315 collects IT service utilization information from target servers 310 using WMI, SNMP or WS-MAN client interfaces and feeds it to a data application layer 325 of a monitor 331 .
- an IPMI client interface 320 gathers information from power management appliances 305 and feeds it to the data acquisition layer 325 .
- This is the orthogonal metric and provides the assessment of maximum, minimum and average electrical consumption.
- Monitor 331 is also coupled to a configuration and scripting interface 333 that inputs threshold and alerts to the monitor 331 .
- the monitor refers to the software component that is responsible for monitoring the measurements from the devices obtaining server—side metrics (e.g.
- Main controller 330 can define the polling interval, event receiving mechanisms and device topology and interface for communication.
- Data acquisition layer 325 in turn feeds it to a database 340 .
- Database 340 is coupled in a feedback arrangement with an analysis engine 345 , which applies processed information for user interface (visualization and interaction) layer 335 that is coupled to an application server or Web server 350 .
- the application server or Web server 350 exchanges the processed information, which includes trend analysis, capacity analysis and efficiency metrics, with a Java client or Web browser 355 at the client side.
- the reporting and publishing logic of this framework includes two kinds of information dissemination to either the form of alerts or reports for energy audits.
- the derived metrics from the framework range from capacity data collection 336 to estimates based on rigorous optimization.
- the reporting of the output of the analysis engine 345 can provide data center efficiency metrics including individual server efficiency and relation to the applications that are running in the data center and the environmental context.
- the output also provides a service that the base model of data center power utilization for tracking the trend and adaptive learning. These outputs can be used to optimize power consumption taking into account the desired business services, server hardware and software characteristics and the data center environment parameters, including cooling and ventilating facilities.
- the strength of this framework is a configurable power monitoring logic and retrieval of IT business service dependencies based on heuristic analysis.
- System 400 includes servers 405 and 410 that are coupled to a control and analysis device 430 that is accessible by a user via a client interface 435 .
- Servers 405 and 410 may be computing devices that implement a virtual machine environment such as VMWARE or XEN.
- a provider 407 412
- Servers 405 and 410 are also connected with power management devices 420 and 425 , respectively.
- Each server 405 ( 410 ) has a respective power management endpoint appliance 415 ( 417 ) that monitors all the running processes (computer power) and their resource utilization at the application level in terms of memory, disk and network usage. This is achieved by monitoring the hypervisor APIs supplied by the virtualization server for hardware resources.
- Device 430 collects this information from devices 405 and 410 and additional information from devices 420 and 425 .
- an energy consumption metric is normalized to the power required to operate the core services like operating systems, network, memory, storage and infrastructure.
- application load may be distributed intelligently among the different physical and logical bins to minimize the total electrical power utilization in terms of computing and cooling.
- FIG. 5 shows the stages involved in the power optimization method, which analyzes the present operating conditions in the services context to compute the operational computing and electric power profiles for effective electric power utilization for any given computing load. Decision rules may be established according to the criticality of business services, which is configurable by the IT administrators and infrastructure providers. For example, the administrator could decide to place an absolute limit on the aggregate rate of power consumption in the data center and identify certain applications whose execution is to be deferred if the power consumption limit has been met. The output of the stage will be delivered to the control logic stage that allocates the application of it among the service in order to minimize total power utilization.
- FIG. 5 a top-level flowchart of the inventive method is shown.
- Environmental, application server power data is input to a filter 505 .
- filter 505 is preferably implemented at the front end of device 230 ( 430 ).
- Filter 505 provides the capability to prioritize information related to electrical power consumption to compute the CPU power metrics that are adequate to correlate with the power data received from, for example, devices 220 and 225 in FIG. 2 .
- the filter supplies the short-term details as well as the long-term trends from the data collected across many end nodes and aggregation devices.
- a function of the filter 505 is to accumulate such resource utilization into a representation that facilitates mapping between the computational load and electrical power. In a smaller data center with especially low level of complexity, the filter may not be required, and the resource utilization information be used as is.
- Filtered data is then input into behavior modeling module 510 .
- Workload or business service characterization is performed in advance for common business services and common server platforms to create basic model for data center power usage. This is stored in knowledge base 515 , providing a second input to module 510 .
- the base modeling of combined characteristics of server and application combinations also helps to reduce the data needed to process the electrical power utilization at any given time of server operation.
- the workload characteristics were normalized to the hardware parameters acquired from the asset database.
- a third input to module 510 is database 520 which provides the current dynamic model information in parameters.
- Module 510 sends database 520 the information about the current datacenter configuration and computational loads which is used to update the database if necessary.
- the current state information is also passed to power metric computation module 525 which uses it to compute the power usage and environmental metrics (e.g. ambient temperature) predicted by the dynamic model.
- the power consumption measure module 530 gets information about the existing distribution of application loads among the various servers from module 510 and also gets predicted power and environment metrics from module 525 .
- Module 530 acquires information on the actual electrical power and thermal and other environmental conditions to the monitoring inputs from, for example, devices 220 and 225 in a time adjusted and correlated manner. If the predicted data is within defined tolerance limits of the actual data, the dynamic model is validated. If not, the information is output to a trend analysis module 535 which provides feedback for updating and refining the dynamic model and algorithm parameters in database 520 .
- the dynamic model is applied to the current application to determine if a reallocation of that load among the servers is required for power usage optimization.
- a better approach of the preferred embodiment develops heuristic rules using a process similar to the back propagation learning algorithms used in neural networks. Neural networks are similar to artificial intelligence methods, trying to model the human learning and intelligence using computer algorithms with a set of neurons and interconnections between themselves based on the learning. In the event that the analysis concludes that a reallocation of application loads or adjustment to the environmental condition infrastructure is required, appropriate instructions are sent to 1) heating and cooling control 540 and 2) server and application load balance control 545 in order to implement the desired changes.
- the major components in the framework for mining the environment for energy conservation include the main controller, monitor, database and analysis engine.
- the amount of data collected through monitor engine may become unmanageably large and may not be fully utilized for energy data collection.
- the framework provides an intelligent configuration module that adapts the customer environment after initial run and facilitates capture and analysis of only selected but relevant information pertaining to the electrical and computing energy calculations.
- the framework is adapted to collect the measurements at the necessary intervals with precision and sampling rate. This module is designed in such a way that the business services and its impact are easily supplemented to the acquisition and analysis logic.
- the overall framework either includes logic to gather asset information of the environment or could be integrated to an external repository. This information is used in the heuristic information for adaptively tracking the association between the electrical and computing power entities.
- the framework gathers the services and the infrastructure either through the dynamic learning or through a static import. This information is then used as the base for analysis of the services in addition to the typical environment behavior. This increases efficiency in terms of deployment effort and automated learning thereafter.
- the synchronization of all individual components in the framework is configurable to an internal time server or to an external network time service.
- the approach described creates global profiles for electrical and computing utilization by systematic processing of power metrics collected using multiple mechanisms for configuration and characterization of various power utilization heuristics.
- the operational profiles specify the amount of computing load at each server, the location of the server to the host that the computing load and the electrical units delivered to the cooling unit on required schedules.
- the method models energy consumption at the process level to maximize the accuracy of power optimization across the entire environment.
- the Web services-based approach works well for distributed enterprises where the communication framework is extended hierarchically through multiple levels of aggregation control to gather the dynamic load conditions and load-balancing operations at locations remote from the central power management system.
- FIG. 5 is illustrative only and that other program entry and exit points, timeout functions, error checking routines and the like (not shown) would normally be implemented in a typical system software. It is also understood that the system software can be implemented to run continuously in an embedded system. Accordingly, start blocks and end blocks are intended to indicate logical beginning and ending points of a portion of code that can be integrated into the main program and called as needed to support continuous system operation. Implementation of these aspects of the invention is readily apparent and well within the grasp of those skilled in the art based on the disclosure herein.
- Computers commonly operate in a networked environment using logical connections to one or more computers.
- the computers used in conjunction with the method may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above.
- the connections include but are not limited to local area network (LAN), wide area network (WAN) and other such networking environments that are commonplace in offices, enterprise wide computer networks, intranets, and the Internet. It will be appreciated that the network connections shown are exemplary and other means of establishing communications links between the computers may be used.
- the computer or computing device typically includes a processor.
- the processor typically includes a CPU such as a microprocessor.
- a CPU generally includes an arithmetic logic unit ALU which performs arithmetic and logical operations, and a control unit, which extracts instructions (e.g. code) from memory and decodes and executes them, calling on the ALU as necessary.
- Memory refers to one or more memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), or electrically erasable programmable read-only memory (EEPROM) chips, by way of further non-limiting example only.
- RAM random access memory
- ROM read-only memory
- PROM programmable read-only memory
- EPROM erasable programmable read-only memory
- EEPROM electrically erasable programmable read-only memory
- ACPI is, at its most basic, a method for describing hardware interfaces in terms abstract enough to allow flexible hardware implementations while allowing shrink-wrap OS code to use such hardware interfaces.
- Operating System-directed Power Management (OSPM) is a model of power (and system) management in which the OS plays a central role and uses global information to optimize system behavior for the task at hand.
- ACPI Hardware is computer hardware with the features necessary to support OSPM and with the interfaces to those features described using the ACPI Description Tables.
- the operating system interacts with ACPI hardware by selecting various ACPI defined states—a mode of operation of either a system, processor or device which defines certain behaviors of the system, process or device.
- ACPI defined states a mode of operation of either a system, processor or device which defines certain behaviors of the system, process or device.
- An incomplete list of states is found below. These various states are for described in, for example, Advanced Configuration and Power Interface Specification (Revision 3.0b Oct. 10, 2006).
- C states are processor power consumption and thermal management states.
- the C 0 state is baseline—the processor in this state executes instructions.
- the C 1 state has the lowest latency.
- the hardware latency in this state must be low enough that the operating software does not consider the latency aspect of the state when deciding whether to use it. Aside from putting the processor in a non-executing power state, this state has no other software-visible effects.
- the C 2 state offers improved power savings over the C 1 state.
- the worst-case hardware latency for this state is provided via the ACPI system firmware and the operating software can use this information to determine when the C 1 state should be used instead of the C 2 state.
- the transition time from C- 2 to C- 0 is on the order of magnitude of 100 nanoseconds, and the processes achieved on a processor level by gating the processor core clock and platform I/O buffers. Aside from putting the processor in a non-executing power state, this state has no other software-visible effects.
- the C 3 state offers improved power savings over the C 1 and C 2 states.
- the bus clock and PLLs are dated.
- Wakeup time is preferably on the order of 50 ⁇ s.
- the worst-case hardware latency for this state is provided via the ACPI system firmware and the operating software can use this information to determine when the C 2 state should be used instead of the C 3 state.
- the processor's caches While in the C 3 state, the processor's caches maintain state but ignore any snoops.
- the operating software is responsible for ensuring that the caches maintain coherency.
- Px states Device and Processor Performance states are power consumption and capability states within the active/executing state C 0 and D 0 .
- the Px states are briefly defined as follows.
- P 0 While a device or processor is in this state, it uses its maximum performance capability and may consume maximum power.
- P 1 In this performance power state, the performance capability of a device or processor is limited below its maximum and consumes less than maximum power.
- Pn In this performance state, the performance capability of a device or processor is at its minimum level and consumes minimal power while remaining in an active state. State n is a maximum number and is processor or device dependent.
- Dynamic voltage scaling is a power management technique in computer architecture, where the voltage used in a component is increased or decreased, depending upon circumstances. Reducing the voltage means increasing program runtime while reducing power consumption.
- Dynamic frequency scaling also known as CPU throttling is a technique in computer architecture where a processor is run at a less-than-maximum frequency in order to conserve power.
- DVFS dynamic voltage and frequency scaling
- the CPU core voltage, clock rate, or both can be altered to decrease power consumption at the price of potential performance. This can be used to optimize the power-performance trade-off.
- DVFS technology includes dynamic voltage scaling and dynamic frequency scaling.
- Dynamic voltage scaling is a power management technique in computer architecture, where the voltage used in a component is increased or decreased depending on circumstances. Dynamic voltage scaling to decrease voltage is known as undervoltage.
- MOSFET-based digital circuits operate using voltages at circuit nodes to represent logical states. The voltage at these nodes switches between a high-voltage and a low voltage during normal operation-when the inputs to a logic gate transition, the transistors making up that gate toggle the gate's output.
- each node in a circuit is a certain amount of capacitance.
- This capacitance arises from various sources, mainly transistors through a capacitance of diffusion capacitance and wires coupling capacitance. Toggling a voltage at a circuit node requires charging or discharging the capacitance in that node. Since capacitance is related to voltage, the time that the node takes depends on the voltage applied. By applying a higher voltage to device a circuit, the capacitance is higher and discharges more quickly, resulting in a faster operation of the circuit and outing for a higher frequency operation.
- a switching power dissipated by a chip using static CMOS gates is determined by the equation C*V 2 *f, where C is the capacitance being switched per clock cycle, V is the voltage and f is the switching frequency. While this formula is not exact, as modern chips are not implemented using only CMOS technology, it does show that overall power consumption decreases with voltage. Accordingly, dynamic voltage scaling is widely used as part of strategies to manage switching power consumption in battery-powered devices such as cell phones and laptop computers. Low voltage modes are used in conjunction with lower clock frequencies to minimize power consumption associated with such as CPUs and DSPs. Only when significant computational power is needed both voltage and frequency be raised. Note that many peripherals also support low-voltage operational modes.
- Dynamic frequency scaling is a technique where processors run at a less than maximum frequency in order to conserve power. Again, the efficiency of some electrical components, such as the voltage regulators, decreases with increasing temperature, so the power use may increase with temperature. Since increasing power use may increase the temperature, increases in voltage or frequency may increase a system power demand even further than the CMOS formula indicates and vice versa.
- ACPI Advanced Configuration and Power Interface
- servers 205 and 210 a FIG. 2 include ACPI hardware and firmware to allow compliance with ACPI standards, and operating systems that are OSPM capable.
- FIG. 5 a top-level flowchart of one method of the present invention as previously discussed, also includes dynamic frequency and voltage scaling control.
- the particulars of any set of available states among a heterogeneous collection of CPUs within a server and/or among a set of servers such as servers 210 and 215 are included in the knowledgebase 515 , which shall update module 510 to allow algorithms of power metric computation module 525 and power consumption management module 530 to incorporate this information for enhanced power optimizations. Note that these added dimensions of CPU state will increase the computational complexity of these algorithms.
- an application such as the analysis engine shown in FIG.
- This operating system may then either call on a server such as server 210 or 215 or a hypervisor such as hypervisor 407 or 412 .
- the hypervisor then interfaces with the hardware to cause the CPU state change.
- the analysis engine shown in FIG. 5 may issue requests for changes to APCI state in any of a plurality of devices, central processing units, motherboards, or any other APCI compliant hardware.
Abstract
A system and method of increasing the efficiency of overall power utilization in data centers by integrating a power management approach based on a comprehensive, dynamic model of the data center created with integrated environmental and computational power monitoring to correlate power usage with different configurations of business services utilization, with the techniques of CPU level power management.
Description
- This application is a continuation of U.S. patent application Ser. No. 12/603,322, filed on Oct. 21, 2009, which claims priority from U.S. Provisional Patent Application No. 61/107,172, filed Oct. 21, 2008, and is a continuation-in-part of U.S. patent application Ser. No. 12/044,530, entitled, “Environmentally Cognizant Power Management,” filed on Mar. 7, 2008, all of which are assigned to the same assignee, and which are incorporated herein by reference in their entireties.
- The invention relates to power management. In particular, a system and method is presented for optimizing power management of computational and electrical appliances provisioned in data centers and server installations.
- Most business organizations today rely on computing power for their business services, including data analysis, supply chain management, inventory tracking, online transactions and customer support. This computing power comes in the form of Web services, Web portals and other open source and proprietary applications hosted in either leased or owned data centers. These data centers use a significant amount of electrical power, both directly through data center computational appliances and indirectly through associated humidity and thermal conditioners. Recent data shows that almost 50% of power delivered to a server farm is spent on cooling infrastructure, while less than 50% is actually utilized in server consumption. The amount of electrical power use during the computational activity inside the server translates into a thermal load. The amount of electric power spent to maintain the operational temperature is also dependent on server air flow characteristics, the relative location of the server hardware within the rack, and other parameters as described later in this disclosure. Even though there is a direct relationship between the computational power utilized by the data centers and supplied electric power, the factors affecting that relationship are many, and the instrumentation and analysis needed to quantify them to the required precision for effective control is challenging. Existing power control mechanisms do not attempt to correlate such utilization with given electrical supply units, and hence fall short of global optimization of the power utilization in data centers and server installations. The above-disclosed related application describes a systematic procedure and apparatus to achieve such monitoring and control using collaborative server computational power measurement and electrical power units consumed under different environmental operational conditions. This related application describes a method to provide the necessary adaptive learning required to address diverse data center server farms and their infrastructure installations. The heuristics used in the method take into account the server hardware, and their associated requirements and locations inside server rack remote locations within the data centers.
- While a number of methods are described in the above related application, a class of techniques used in low-level power management but which have not been applied in monitoring and controlling the direct relationship between the computational power utilized by the data centers and supplied electric power are described in the present application. Specifically, techniques for controlling the amount of energy consumption at the motherboard and CPU level may also be used in the monitoring and control of power utilized by servers in data centers
- Heretofore, such capabilities have been applied in a piecemeal manner primarily to power management in laptop and other mobile systems. There remains a need in the art to apply this technique to systematic way such as described in the above cross-referenced application, to the balance of the computational power utilized by the data centers versus supplied electric power.
- A system and method of increasing the efficiency of overall power utilization in data centers by integrating a power management approach based on a comprehensive, dynamic model of the data center created with integrated environmental and computational power monitoring to correlate power usage with different configurations of business services utilization with the techniques of CPU level power management
- For a better understanding of the present invention, reference is made to the following description and accompanying drawings, while the scope of the invention is set forth in the appended claims:
-
FIG. 1 illustrates major categories of energy consumption; -
FIG. 2 is an exemplary architecture for power management in a server farm or data center environment according to the present invention. -
FIG. 3 is an exemplary framework related to the architecture shown inFIG. 2 . -
FIG. 4 is an exemplary architecture for power management in a virtual machine environment according to the present invention. -
FIG. 5 is an exemplary top level flowchart of a method according to the present invention. -
FIG. 1 depicts the major sections of energy consumption inserver installations 100.Total energy 110 is a summation of the power used forcomputational purposes 120 and the power used for heating andcooling functions 130. Environmentally based power management uses these major consumption areas to minimize the total energy usage in an adaptive manner. The iterative analysis begins by computing the initial measure of total electrical power consumed by the server and network hardware configuration from the asset information. The assets denote the electrical and computing units in the infrastructure and cover, for example, servers, routers, switches, cooling units, racks, sensors and power distribution panels. This base computation is then supplemented by the electrical power delivered to the cooling and other infrastructure elements. This electrical side power consumption is tracked against the computational power consumed by business services and correlated with real time electrical metering at the server level. The asset information and other configuration information can either be supplied directly to the power management system or imported from existing IT management systems, which is common in an enterprise level installation. The environmentally conscious power management system and method of the present invention increases the efficiency of overall power utilization by using coordinated power monitoring at the electrical power distribution side as well as in the consumption side of server installations using simplified common management interfaces (e.g. simple network management protocol (SNMP) web services) for environmental and server monitoring. The system and method improves the efficiency of power utilization per computational unit and provides policies for dynamic load allocation capabilities based on application loads and environmental contexts, e.g. rack configurations, cooling units and thermal characteristics of the data center room. The dynamic computational load balancing policies of the present invention can be adopted by (i) positioning the servers at environmentally optimized locations, (ii) scheduling the computational tasks at physically diverse locations, and (iii) providing VMOTION (utility for migration of virtual machines between physical servers offered by VMWARE) in the virtual environments leveraging virtualization technology. The proposed approach addresses modifications needed to measure the electrical power requirements in such virtualization environments. - A coordinated framework is the most important ingredient to maximize the benefits of information gathered by the environmental and server measuring agents denoted as monitors in the figures illustrated herewith. Practices address local electrical optimization which might not accurately determine the total requirements. In an environment where multiple business services are consolidated at the server level, they might help to reduce the power consumed at each individual server. However they do not account for the operating conditions and thus become ineffective in situations where those conditions are significant, such as when the servers are configured in a dense arrangement like server blades. In addition to consolidating multiple business services into a single or few servers for minimizing power consumption, it is also important to determine the optimum location of that server hardware based on the environmental context. For example, an application running on a server with a rack mounted cooling unit would likely use less overall power than the same application running on an identical server in a less efficient cooling environment. Similarly, an application running on a server with a lowpower hard drive would likely use less power than the same application running on a server with large power consumption or inefficient heat dissipation design. The selection of servers for load balancing is mostly ad hoc under current practices and does not detailed analysis.
- The most important challenge in power management is the lack of components to correlate the computational processing power related to services and electrical power consumption with the accuracy needed for globally optimized load-balancing and control. In addition to the challenges needed to monitor a single location environment, IT in today's enterprises are typically located at geographically distant locations for better utilization of human, time and material resources. These geographic variables are not taken into account in present electrical power operational profiles due to the lack of a comprehensive solution to address such capabilities. As business services are rendered for multiple geospatial locations, it is critical to include coordination among electrical power management and computational processing operations at the enterprise level within the global context of business services. The optimization of overall power utilization thus does require a coordinated framework, systematic procedure and control elements to distribute the business computing load both at the physical location and logical locations (as applications and server clusters or virtual machines) by configurable and adaptive monitoring and continuous analysis of the environment for global power management decisions.
- The server monitoring tools and appliances employ either agent or non-agent based technologies to measure application load, starting with determining the number of instances of each application running on each server, using different mechanisms primarily depending on the operating systems hosted in server hardware. Typical monitoring interfaces available for monitoring and measurement include Windows management instrumentation (WMI) on Microsoft platforms, simple network management protocol (SNMP), and Web service management (WS-MAN). The granularity of the information necessary to associate the application loads of the electrical power used is not directly computable from the standard interfaces. Thus it is useful to obtain the right metrics from these existing base metrics that are usable in electrical power computations. The environmentally-based power management framework addresses this constraint by aggregating the process level information from different mechanisms (including network computer monitoring devices such as Raritan's COMMAND CENTER NOC, baseboard management controller such as Raritan's KIRA, and power environmental monitoring devices such as Raritan DOMINION PX) to improve the accuracy of the predicted electrical power unit consumption based on environmental factors, server hardware characteristics, operating system overheads and running application to provide business services. As shown below, a first level of power metrics is obtained from server hardware and power supply units installed within the servers and a second level of power metrics is obtained from operating system and applications executed within the servers. A third level of metrics is computed from the live monitoring of actual business utilization, configuration, topology, thermal activity and electrical power usage.
- Another constraint in the power management operations that the server monitoring instrumentation and electrical environment monitoring instrumentation are installed separately and maintained independently, thus making calibration and coordination of these measurements relatively difficult. The environmentally cognizant power management system and method integrates both server and environmental monitoring and provides cooperative collection and processing of metrics, thus improving the scope of optimization at the global level. In order to improve the interoperability of measurement devices and distributed data collection, a common information model (CIM) is proposed, and management interfaces that support this CIM profiles are recommended. Intelligent platform management interface (IPMI) is one such standard and Raritan Baseboard Management Controller (KIRA) and DOMINION PX are a few examples of monitoring and management devices that support the Intelligent Platform management Interface (IPMI) interface making it suitable for power optimization applications.
- The environmentally cognizant power management framework also applies to both virtual and physical server environments. In particular, virtualization technology provides dynamic load balancing operations that help simplify collaborative power management. As described later, the proposed approach can be adopted to a physical server installations as well as virtual server installations.
- In addition to the instrumentation necessary to collect a server computational power metrics and electrical power measurements at desired interfaces during programmed intervals, the synchronization of this information needs to be clearly determined for accurate correlation of events and submission of both utilization supply. The management controller and power optimization (330) system of
FIG. 3 supports the necessary components to order and group the data set collected over multiple interfaces to various time intervals. The information required to compute the long-term trending is different from localized short bursts of measurements. In such cases, the management controller configures the necessary data collection devices to collect the measurements at different sampling intervals between maintenance cycles as necessary. - Referring now to
FIG. 2 , there is shown asystem 200 that uses coordinated power management server installations, data centers and other such constructs.System 200 includesservers monitoring device 215.Servers monitoring device 215 may be Raritan's Command Center NOC or any other similar device.Servers power management devices Power management devices monitoring device 215 andpower management devices 225 are further coupled to a monitoring control andanalysis device 230 that is accessible to the user by aclient interface 235. - The environmentally cognizant power management system and method uses the above management components as building blocks for power management framework. In particular, the management and
monitoring device 215 determines in real-time the operating system and applications running a server. In addition, the management andmonitoring devices 215 determines and real-time the operating system and applications monitoring on each server. In addition, the management andmonitoring devices 215 monitor server hardware parameters using a Baseboard Management Controller's hardware that may be available as embedded modules or parts, such as for example, in Raritan's KIRA device. These parameters, including the power cycle events and CPU temperatures, provide an additional level of power metrics to correlate to the CPU utilization and the computational load created by the business applications.Power management devices - The interaction of these components and power optimization is detailed in
FIG. 2 . In this scenario, the business applications are distributed according to operating profiles computed by the centralized processing engine (main controller) running in control andanalysis device 230. That is, business applications are scheduled to be executed at a particular server based on its efficiency and power and thermal profiles and its position within the rack, location within the server room or facility, as well as environmental conditions. The processing logic indevice 230 gathers environmental conditions from the environmental sensors which are coupled to a common monitor (or integrated intopower management devices 220 and 225) and the measurement of electric power usage that are directly available frompower management devices -
- First, the current draw and efficiency of the power supply units which convert alternating current (AC) power into direct current (DC) power used inside the server hardware. This information is acquired from the configuration management database (CMDB.), known as asset data in IT terminology. The asset information includes the nameplate data that describes the maximum power drawn by the hardware and the inlet and outlet temperatures necessary for proper operating conditions based on its physical design and functional characteristics. In addition to the component level information, the physical location of the racks, the relative distance of servers from the cooling systems and orientation of racks are also fed into the system either from the CMDB, if available, or from other data sources.
- Second, the base level power requirements for the operating systems and applications running on servers. This is dependent on the utilization of central processing unit or CPU, random access memory, hard disk, and other resources in addition to the operating system implementation. This information is typically imported from the IT infrastructure systems or automatically collected through the Baseboard Management Controller (BMC).
- Third, the actual utilization of business computing services within the environmental and chronological context. These process level metrics are obtained by monitoring
appliance 215 through WMI and SNMP interfaces at periodic intervals as computed from the main controller.
- In order to improve the accuracy of correlating electrical power requirements with the information from various appliances, the measurement data are synchronized to a clock server that runs inside
device 230. The sequencing operation of the measurement data acquired from any device is validated based on the time of arrival of the relative event occurrences before storage for further processing. In addition to the synchronization capabilities, the Web services capability automates the discovery and communication between the information provider services and the processing engine which runs indevice 230. Before the start of monitoring and data collection, all measurement devices must be discovered and properly registered in order to successfully interpret the events with its context. The context of the operating conditions is important to the measurement. For example, the electrical power usage of computer room air conditioner (CRAG) units in a lightly loaded environment is to be treated differently from the measurement of highly loaded operating conditions. The data received throughout the entire system is thus interpreted within the operational context of the global level. The monitor data are now accessible to receivers, which, in the usual scenario, are provided in the main controller in the system. If there is a failure to receive data from one device or multiple devices, the management controller adapts the next level metrics or derived metrics from previous context or earlier data acquisition until the system is back to normal operation. The standard interface of using Web services fit well into scalable reconfigurable enterprise platform and the proposed approach can be easily adopted into its existing infrastructure. The prior art power management for information technology services does not provide the capabilities of coordinated operations over the network and thus cannot benefit from load-balancing across infrastructure and computational equipment in power utilization. The inventive system and method address this constraint by provisioning network aware services in each appliance. This facilitates real-time synchronization between different units in the migration processing across multiple units. A network enabled Web services technology supports both reactive and proactive power monitoring, which is valuable to complex environments where the behavior of applications and short term failures or irreparable crashes. Web services can provide configuration updates and enable reconfiguration by subscription methods between peer-to-peer systems, which in the present case are the data measurement devices and data receivers for processing the data. - Where the power (and hence cooling) is of critical importance, server level power monitoring is highly beneficial for information technology management, specifically for auditing, capacity planning, thermal assessment, and performance improvement. These scenarios essentially require power usage and active monitoring on servers to determine the need for rescheduling the load based on predefined and/or dynamic power policies. The dynamic power management policies are either stored in the system or referred from an external network management system. Basic level of power profiles are supported at the operating system level and advanced configuration and power interface (ACPI) provides a way to control this aspect. The proposed approach applies to any power profiles including ACPI and uses appropriate metrics for different kinds of profiles. Current power policies that utilized ACPI focus on reliable power feed to the computing units based on the availability of electrical power from alternate sources like uninterruptible power supply generators.
- Referring now to
FIG. 3 , there is shown a framework of an intelligentpower monitoring system 300 for server farms or data centers in accordance with the invention. In afirst phase 301,system 300 includes apower management appliance 305 that acts as a power distribution unit (PDU) to targetservers 310 and enables users to measure power at the outlet level, providing an accurate view of power consumption at atarget server 310 or a PDU level (total power distributed). Most of the existing power distribution units measure rack level or branch circuit level power measurements, which might not be sufficient for power monitoring at the server level. In asecond phase 302, management andmonitoring device 315 collects IT service utilization information fromtarget servers 310 using WMI, SNMP or WS-MAN client interfaces and feeds it to adata application layer 325 of amonitor 331. In addition, an IPMI client interface 320 gathers information frompower management appliances 305 and feeds it to thedata acquisition layer 325. This is the orthogonal metric and provides the assessment of maximum, minimum and average electrical consumption.Monitor 331 is also coupled to a configuration andscripting interface 333 that inputs threshold and alerts to themonitor 331. In this description, the monitor refers to the software component that is responsible for monitoring the measurements from the devices obtaining server—side metrics (e.g. operating systems, CPU temperature, application instances) as well as those obtaining the electrical side metrics (e.g. power usage, temperature, air flow and humidity). This is the data interface layer of the system, and is configurable from themain controller 330.Main controller 330 can define the polling interval, event receiving mechanisms and device topology and interface for communication. -
Data acquisition layer 325 in turn feeds it to adatabase 340.Database 340 is coupled in a feedback arrangement with ananalysis engine 345, which applies processed information for user interface (visualization and interaction)layer 335 that is coupled to an application server orWeb server 350. In athird phase 303, the application server orWeb server 350 exchanges the processed information, which includes trend analysis, capacity analysis and efficiency metrics, with a Java client orWeb browser 355 at the client side. - The reporting and publishing logic of this framework includes two kinds of information dissemination to either the form of alerts or reports for energy audits. As depicted in
FIG. 3 , the derived metrics from the framework range fromcapacity data collection 336 to estimates based on rigorous optimization. The reporting of the output of theanalysis engine 345 can provide data center efficiency metrics including individual server efficiency and relation to the applications that are running in the data center and the environmental context. The output also provides a service that the base model of data center power utilization for tracking the trend and adaptive learning. These outputs can be used to optimize power consumption taking into account the desired business services, server hardware and software characteristics and the data center environment parameters, including cooling and ventilating facilities. The strength of this framework is a configurable power monitoring logic and retrieval of IT business service dependencies based on heuristic analysis. - Referring now to
FIG. 4 , there is shown asystem 400 that uses coordinated power management in a virtual machine implementation.System 400 includesservers client interface 435.Servers Servers power management devices devices devices - In both scenarios illustrated in
FIGS. 2 and 4 , an energy consumption metric is normalized to the power required to operate the core services like operating systems, network, memory, storage and infrastructure. In an exemplary embodiment, application load may be distributed intelligently among the different physical and logical bins to minimize the total electrical power utilization in terms of computing and cooling. - The data driven framework facilitates monitoring and provides interface for dynamic control of electrical and computing power management. This method reduces the cost of equivalent computing power needed by providing the basis for an optimal distribution of total computing load needed by the business at any given time. In general,
FIG. 5 shows the stages involved in the power optimization method, which analyzes the present operating conditions in the services context to compute the operational computing and electric power profiles for effective electric power utilization for any given computing load. Decision rules may be established according to the criticality of business services, which is configurable by the IT administrators and infrastructure providers. For example, the administrator could decide to place an absolute limit on the aggregate rate of power consumption in the data center and identify certain applications whose execution is to be deferred if the power consumption limit has been met. The output of the stage will be delivered to the control logic stage that allocates the application of it among the service in order to minimize total power utilization. - Referring specifically to
FIG. 5 , a top-level flowchart of the inventive method is shown. Environmental, application server power data is input to afilter 505. Since the volume of data to be collected throughdevice 230 inFIG. 2 and device 430 inFIG. 4 is quite large,filter 505 is preferably implemented at the front end of device 230 (430).Filter 505 provides the capability to prioritize information related to electrical power consumption to compute the CPU power metrics that are adequate to correlate with the power data received from, for example,devices FIG. 2 . The filter supplies the short-term details as well as the long-term trends from the data collected across many end nodes and aggregation devices. The CPU usage, main memory, disk I/O, and the CPU time slice shared among many virtual machines are examples of metrics that have major impact on the electrical power consumption on that server hardware. A function of thefilter 505 is to accumulate such resource utilization into a representation that facilitates mapping between the computational load and electrical power. In a smaller data center with especially low level of complexity, the filter may not be required, and the resource utilization information be used as is. - Filtered data is then input into
behavior modeling module 510. Workload or business service characterization is performed in advance for common business services and common server platforms to create basic model for data center power usage. This is stored inknowledge base 515, providing a second input tomodule 510. The base modeling of combined characteristics of server and application combinations also helps to reduce the data needed to process the electrical power utilization at any given time of server operation. The workload characteristics were normalized to the hardware parameters acquired from the asset database. A third input tomodule 510 isdatabase 520 which provides the current dynamic model information in parameters. -
Module 510 sendsdatabase 520 the information about the current datacenter configuration and computational loads which is used to update the database if necessary. The current state information is also passed to powermetric computation module 525 which uses it to compute the power usage and environmental metrics (e.g. ambient temperature) predicted by the dynamic model. The powerconsumption measure module 530 gets information about the existing distribution of application loads among the various servers frommodule 510 and also gets predicted power and environment metrics frommodule 525.Module 530 acquires information on the actual electrical power and thermal and other environmental conditions to the monitoring inputs from, for example,devices trend analysis module 535 which provides feedback for updating and refining the dynamic model and algorithm parameters indatabase 520. - As a final step, the dynamic model is applied to the current application to determine if a reallocation of that load among the servers is required for power usage optimization. Given the models to predict power usage based on the application of common environmental factors and server hardware and software characteristics, there are numerous methods for making that determination known in the art. For example, a simple approach would be to rank order the servers based on the predicted incremental power draw for an increment of computational load under existing environmental conditions and rank order the applications based on the anticipated schedule of computational load. Starting with the application with a large load this process would allocate instances of the applications first to the most efficient server and, as a server's maximum computational load was reached, proceeding to the next most efficient server until done. This approach will yield reasonable results and has the advantage of being simple and fast, it will often result in a sub optimal allocation. A better approach of the preferred embodiment develops heuristic rules using a process similar to the back propagation learning algorithms used in neural networks. Neural networks are similar to artificial intelligence methods, trying to model the human learning and intelligence using computer algorithms with a set of neurons and interconnections between themselves based on the learning. In the event that the analysis concludes that a reallocation of application loads or adjustment to the environmental condition infrastructure is required, appropriate instructions are sent to 1) heating and
cooling control 540 and 2) server and applicationload balance control 545 in order to implement the desired changes. - In general the major components in the framework for mining the environment for energy conservation include the main controller, monitor, database and analysis engine. The amount of data collected through monitor engine may become unmanageably large and may not be fully utilized for energy data collection. The framework provides an intelligent configuration module that adapts the customer environment after initial run and facilitates capture and analysis of only selected but relevant information pertaining to the electrical and computing energy calculations. The framework is adapted to collect the measurements at the necessary intervals with precision and sampling rate. This module is designed in such a way that the business services and its impact are easily supplemented to the acquisition and analysis logic. The overall framework either includes logic to gather asset information of the environment or could be integrated to an external repository. This information is used in the heuristic information for adaptively tracking the association between the electrical and computing power entities. The framework gathers the services and the infrastructure either through the dynamic learning or through a static import. This information is then used as the base for analysis of the services in addition to the typical environment behavior. This increases efficiency in terms of deployment effort and automated learning thereafter. The synchronization of all individual components in the framework is configurable to an internal time server or to an external network time service.
- The approach described creates global profiles for electrical and computing utilization by systematic processing of power metrics collected using multiple mechanisms for configuration and characterization of various power utilization heuristics. The operational profiles specify the amount of computing load at each server, the location of the server to the host that the computing load and the electrical units delivered to the cooling unit on required schedules. The method models energy consumption at the process level to maximize the accuracy of power optimization across the entire environment. The Web services-based approach works well for distributed enterprises where the communication framework is extended hierarchically through multiple levels of aggregation control to gather the dynamic load conditions and load-balancing operations at locations remote from the central power management system.
- It is understood that
FIG. 5 is illustrative only and that other program entry and exit points, timeout functions, error checking routines and the like (not shown) would normally be implemented in a typical system software. It is also understood that the system software can be implemented to run continuously in an embedded system. Accordingly, start blocks and end blocks are intended to indicate logical beginning and ending points of a portion of code that can be integrated into the main program and called as needed to support continuous system operation. Implementation of these aspects of the invention is readily apparent and well within the grasp of those skilled in the art based on the disclosure herein. - Although an exemplary network environment as described above, in the network of interconnected computers, servers, applications and other devices are applicable and can be used with respect to the method described above. Computers commonly operate in a networked environment using logical connections to one or more computers. The computers used in conjunction with the method may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above. The connections include but are not limited to local area network (LAN), wide area network (WAN) and other such networking environments that are commonplace in offices, enterprise wide computer networks, intranets, and the Internet. It will be appreciated that the network connections shown are exemplary and other means of establishing communications links between the computers may be used. The purpose of illustration, programs and other executable program components such as the operating system are illustrated herein as discrete blocks, although it is recognized that such programs and components reside at various times in different storage components of the computer, and are executed by the data processor(s) of the computer. Different combinations of hardware and software can be used to carry out the teachings of the present invention. The computer or computing device typically includes a processor. The processor typically includes a CPU such as a microprocessor. A CPU generally includes an arithmetic logic unit ALU which performs arithmetic and logical operations, and a control unit, which extracts instructions (e.g. code) from memory and decodes and executes them, calling on the ALU as necessary. Memory as used herein refers to one or more memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), or electrically erasable programmable read-only memory (EEPROM) chips, by way of further non-limiting example only. Memory may be internal or external to an integrated unit including a processor. Memory preferably stores a computer program, e.g., code or sequence of instructions being operable by a processor.
- As previously noted, in the event that analysis concludes that a reallocation of application loads or an adjustment to the environmental conditioning infrastructure is required, appropriate instructions are sent to 1) heating and
cooling control 540 and 2) server and applicationload balancing control 545 in order to implement the desired changes. APCI enabled Hardware and operating systems allow another method of reducing energy costs via efficient energy consumption. Specifically, techniques for controlling the amount of energy consumption at the device, motherboard, chip set and CPU level may also be used in addition tosteps - The operating system interacts with ACPI hardware by selecting various ACPI defined states—a mode of operation of either a system, processor or device which defines certain behaviors of the system, process or device. An incomplete list of states is found below. These various states are for described in, for example, Advanced Configuration and Power Interface Specification (Revision 3.0b Oct. 10, 2006).
-
- Global System States which represent an entire system state which is visible to use, and divided into four states G0, G1, G2, G3;
- Sleeping States residing within global system state G1 (except for S5) and divided into five states 51, S2, S3, S4, S5;
- Device Power states usually invisible to the user divided into four states—D0, D1, D2, and D3 and which vary with each device;
- CPU Power states also known as CPU sleep states which reside within the global system state G) and are divided into five states C0, C1, C2, C3, and C4;
- CPU/Device Performance states which control CPU/Device power management when it is still active, including usually clock speed and voltage variance depending on workload, and which will have a CPI dependent number of states; and
- CPU Thermal Monitor which throttles the CPU to a low performance state when temperature exceeds a threshold, wherein TM1 throttling is done by changing the duty cycle and in TM2, throttling is done by changing the clock speed or core voltage (P state).
- For purposes of power management and determining the amount of electrical power used by a CPU or device, obviously establishing a G1 state or sleeping state is useful, with S1-S4 determining wake up latency of 2 seconds to 30 seconds. Further, various CPU power states (C-states) also allow useful methods to have an operating system regulate power at the CPU levels. C states are processor power consumption and thermal management states. The C0 state is baseline—the processor in this state executes instructions. The C1 state has the lowest latency. The hardware latency in this state must be low enough that the operating software does not consider the latency aspect of the state when deciding whether to use it. Aside from putting the processor in a non-executing power state, this state has no other software-visible effects. This is done in actual execution by using the assembly instruction “halt” with a wake-up time having an order of magnitude of 10 ns. The C2 state offers improved power savings over the C1 state. The worst-case hardware latency for this state is provided via the ACPI system firmware and the operating software can use this information to determine when the C1 state should be used instead of the C2 state. The transition time from C-2 to C-0 is on the order of magnitude of 100 nanoseconds, and the processes achieved on a processor level by gating the processor core clock and platform I/O buffers. Aside from putting the processor in a non-executing power state, this state has no other software-visible effects. The C3 state offers improved power savings over the C1 and C2 states. In the C-3 state, the bus clock and PLLs are dated. Wakeup time is preferably on the order of 50 μs. The worst-case hardware latency for this state is provided via the ACPI system firmware and the operating software can use this information to determine when the C2 state should be used instead of the C3 state. While in the C3 state, the processor's caches maintain state but ignore any snoops. The operating software is responsible for ensuring that the caches maintain coherency.
- Even further, Device and Processor Performance states (Px states) are power consumption and capability states within the active/executing state C0 and D0. The Px states are briefly defined as follows. P0—While a device or processor is in this state, it uses its maximum performance capability and may consume maximum power. P1—In this performance power state, the performance capability of a device or processor is limited below its maximum and consumes less than maximum power. Pn—In this performance state, the performance capability of a device or processor is at its minimum level and consumes minimal power while remaining in an active state. State n is a maximum number and is processor or device dependent.
- The P-states can be achieved by dynamic voltage scaling and dynamic frequency scaling. Dynamic voltage scaling is a power management technique in computer architecture, where the voltage used in a component is increased or decreased, depending upon circumstances. Reducing the voltage means increasing program runtime while reducing power consumption. Dynamic frequency scaling (also known as CPU throttling) is a technique in computer architecture where a processor is run at a less-than-maximum frequency in order to conserve power.
- In this technique of dynamic voltage and frequency scaling (DVFS), the CPU core voltage, clock rate, or both, can be altered to decrease power consumption at the price of potential performance. This can be used to optimize the power-performance trade-off. DVFS technology includes dynamic voltage scaling and dynamic frequency scaling. Dynamic voltage scaling is a power management technique in computer architecture, where the voltage used in a component is increased or decreased depending on circumstances. Dynamic voltage scaling to decrease voltage is known as undervoltage. MOSFET-based digital circuits operate using voltages at circuit nodes to represent logical states. The voltage at these nodes switches between a high-voltage and a low voltage during normal operation-when the inputs to a logic gate transition, the transistors making up that gate toggle the gate's output. At each node in a circuit is a certain amount of capacitance. This capacitance arises from various sources, mainly transistors through a capacitance of diffusion capacitance and wires coupling capacitance. Toggling a voltage at a circuit node requires charging or discharging the capacitance in that node. Since capacitance is related to voltage, the time that the node takes depends on the voltage applied. By applying a higher voltage to device a circuit, the capacitance is higher and discharges more quickly, resulting in a faster operation of the circuit and outing for a higher frequency operation. A switching power dissipated by a chip using static CMOS gates is determined by the equation C*V2*f, where C is the capacitance being switched per clock cycle, V is the voltage and f is the switching frequency. While this formula is not exact, as modern chips are not implemented using only CMOS technology, it does show that overall power consumption decreases with voltage. Accordingly, dynamic voltage scaling is widely used as part of strategies to manage switching power consumption in battery-powered devices such as cell phones and laptop computers. Low voltage modes are used in conjunction with lower clock frequencies to minimize power consumption associated with such as CPUs and DSPs. Only when significant computational power is needed both voltage and frequency be raised. Note that many peripherals also support low-voltage operational modes. Please note that reducing the voltage means that circuit switch is slower, producing the maximum frequency at which that's circuit can run. This in turn, reduces the rate at which program instructions can be issued, which may increase runtime for program segments which are succeeded significantly CPU-bound. Thus there is a complex trade-off between power management and computational load. Also note that the efficiency of some logical components, such as the regulators, decreases with increasing temperature, so the power use may increase with temperature. Since increasing power use may increase the temperature, increasing the voltage or frequency may also increase system power demands even faster than the above CMOS formula indicates.
- Dynamic frequency scaling is a technique where processors run at a less than maximum frequency in order to conserve power. Again, the efficiency of some electrical components, such as the voltage regulators, decreases with increasing temperature, so the power use may increase with temperature. Since increasing power use may increase the temperature, increases in voltage or frequency may increase a system power demand even further than the CMOS formula indicates and vice versa.
- Note that each of the above techniques, there is a reinforcing effect, where the production of either voltage or frequency applied to the CPU as whole or functional units within the CPU also reduces the amount of heat generated by the CPU or its components. As such, less cooling is the necessary for the motherboard, further reducing overall system energy consumption.
- The above described techniques are well known in the art of power management for laptops and other such devices. These techniques have been caught codified in a number of ways, again most prominently the Advanced Configuration and Power Interface (ACPI). ACPI is an open industry specification within an established industry. Implementation of these aspects of the invention is readily apparent and well within the grasp of those skilled in the art based on the disclosure herein.
- In one embodiment of the present invention,
servers 205 and 210 aFIG. 2 include ACPI hardware and firmware to allow compliance with ACPI standards, and operating systems that are OSPM capable.FIG. 5 , a top-level flowchart of one method of the present invention as previously discussed, also includes dynamic frequency and voltage scaling control. The particulars of any set of available states among a heterogeneous collection of CPUs within a server and/or among a set of servers such asservers knowledgebase 515, which shall updatemodule 510 to allow algorithms of powermetric computation module 525 and powerconsumption management module 530 to incorporate this information for enhanced power optimizations. Note that these added dimensions of CPU state will increase the computational complexity of these algorithms. In operation, an application such as the analysis engine shown inFIG. 5 will issue an APCI call to an operating system to change the CPU state. This operating system may then either call on a server such asserver hypervisor - In another embodiment of the present invention, the analysis engine shown in
FIG. 5 may issue requests for changes to APCI state in any of a plurality of devices, central processing units, motherboards, or any other APCI compliant hardware. - While the foregoing description and drawings represent the preferred embodiments of the present invention, it will be understood that various changes and modifications may be made without departing from the spirit and scope of the present invention.
Claims (10)
1. A method for cognizant power management in a distributed computing system having a plurality of processing units, comprising the steps of:
generating a predictive behavior power utilization model based on first values of process level information associated with the processing units and first values of environmental metrics for the computing system;
predicting a power utilization of the computing system according to the predictive behavior power utilization model and a selected computational load to be distributed among the processing units;
distributing the computational load to the processing units computing system based on based on the predictive behavior power utilization model;
determining an actual power utilization by gathering second values for the process level information and the environmental metrics; and
updating the predictive behavior power utilization model according to actual power utilization when a difference between the actual power utilization and the predicted power utilization exceeds a predetermined threshold.
2. The method of claim 1 , wherein the plurality of processing units includes at least one of a plurality of central processing units, a plurality of functional units within central processing units, and a plurality of hypervisors managing virtual machines.
3. The method of claim 2 , further comprising the step of:
setting a state in at least one of the plurality of processing units based on the predictive behavior power utilization model before carrying out the distributing step, wherein the state is one of a plurality of performance states.
4. The method of claim 3 , wherein setting one of a plurality of performance states includes changing the frequency of a processing unit.
5. The method of claim 3 , wherein setting one of a plurality of performance states includes changing the voltage of a processing unit.
6. The method of claim 3 , wherein setting one of a plurality of performance states includes changing the frequency and the voltage of a processing unit.
7. The method of claim 3 , wherein one of a plurality of performance states includes at least a first performance state and a second performance state, wherein the first performance state can bear a greater computational load than the second state.
8. The method of claim 3 , wherein the state is one of a plurality of sleep states.
9. The method of claim 8 , wherein one of a plurality of sleep states includes at least a first sleep state and second sleep state, wherein the first sleep state has a lower wake up time than the second sleep state.
10. The method of claim 1 , further comprising the step of:
gathering the first values of the process level information from the plurality of processing and the first values of environmental metrics for the computing system before carrying out the predictive step.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/790,601 US20130191658A1 (en) | 2008-03-07 | 2013-03-08 | Methods of achieving cognizant power management |
US14/877,621 US10289184B2 (en) | 2008-03-07 | 2015-10-07 | Methods of achieving cognizant power management |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US12/044,530 US8671294B2 (en) | 2008-03-07 | 2008-03-07 | Environmentally cognizant power management |
US10717208P | 2008-10-21 | 2008-10-21 | |
US12/603,322 US8429431B2 (en) | 2008-03-07 | 2009-10-21 | Methods of achieving cognizant power management |
US13/790,601 US20130191658A1 (en) | 2008-03-07 | 2013-03-08 | Methods of achieving cognizant power management |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/603,322 Continuation US8429431B2 (en) | 2008-03-07 | 2009-10-21 | Methods of achieving cognizant power management |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/877,621 Continuation US10289184B2 (en) | 2008-03-07 | 2015-10-07 | Methods of achieving cognizant power management |
Publications (1)
Publication Number | Publication Date |
---|---|
US20130191658A1 true US20130191658A1 (en) | 2013-07-25 |
Family
ID=42119660
Family Applications (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/603,322 Active 2028-12-12 US8429431B2 (en) | 2008-03-07 | 2009-10-21 | Methods of achieving cognizant power management |
US13/790,601 Abandoned US20130191658A1 (en) | 2008-03-07 | 2013-03-08 | Methods of achieving cognizant power management |
US14/877,621 Active US10289184B2 (en) | 2008-03-07 | 2015-10-07 | Methods of achieving cognizant power management |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US12/603,322 Active 2028-12-12 US8429431B2 (en) | 2008-03-07 | 2009-10-21 | Methods of achieving cognizant power management |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/877,621 Active US10289184B2 (en) | 2008-03-07 | 2015-10-07 | Methods of achieving cognizant power management |
Country Status (7)
Country | Link |
---|---|
US (3) | US8429431B2 (en) |
EP (1) | EP2350770A4 (en) |
JP (1) | JP2012506597A (en) |
CN (1) | CN102227693A (en) |
AU (1) | AU2009308467A1 (en) |
CA (1) | CA2741088C (en) |
WO (1) | WO2010048316A1 (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120078429A1 (en) * | 2010-09-24 | 2012-03-29 | Patrick Edward Weston | APPARATUS AND METHOD FOR COLLECTING AND DISTRIBUTING POWER USAGE DATA FROM RACK POWER DISTRIBUTION UNITS (RPDUs) USING A WIRELESS SENSOR NETWORK |
US20130151696A1 (en) * | 2011-12-12 | 2013-06-13 | Delta Electronics, Inc. | Trigger method of computational procedure for virtual maching migration and application program for the same |
US8811377B1 (en) | 2010-08-30 | 2014-08-19 | Synapsense Corporation | Apparatus and method for instrumenting devices to measure power usage using a multi-tier wireless network |
US20140365641A1 (en) * | 2013-06-11 | 2014-12-11 | Samsung Electronics Co., Ltd | Processor module, server system and method of controlling processor module |
US9154367B1 (en) * | 2011-12-27 | 2015-10-06 | Google Inc. | Load balancing and content preservation |
WO2016021986A1 (en) * | 2014-08-07 | 2016-02-11 | Samsung Electronics Co., Ltd. | Method and apparatus for executing software in electronic device |
US9696781B2 (en) * | 2015-05-28 | 2017-07-04 | Cisco Technology, Inc. | Automated power control for reducing power usage in communications networks |
WO2020246985A1 (en) * | 2019-06-07 | 2020-12-10 | Hewlett-Packard Development Company, L.P. | Device operation modifications based on monitored power levels |
US11269677B2 (en) | 2019-08-08 | 2022-03-08 | Vigyanlabs Innovations Private Limited | System and method to analyze and optimize application level resource and energy consumption by data center servers |
Families Citing this family (67)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9021074B2 (en) * | 2007-11-08 | 2015-04-28 | International Business Machines Corporation | System and method for providing server status awareness |
US8429431B2 (en) * | 2008-03-07 | 2013-04-23 | Raritan Americas, Inc. | Methods of achieving cognizant power management |
JP4724730B2 (en) * | 2008-04-09 | 2011-07-13 | 株式会社日立製作所 | Information processing system operation management method, operation management program, operation management apparatus, and information processing system |
US20100161368A1 (en) * | 2008-12-23 | 2010-06-24 | International Business Machines Corporation | Managing energy in a data center |
US9239994B2 (en) * | 2009-02-25 | 2016-01-19 | Empire Technology Development Llc | Data centers task mapping |
DE102009050170B4 (en) * | 2009-10-21 | 2013-08-01 | Diehl Ako Stiftung & Co. Kg | Home automation and home information system |
US9235250B2 (en) * | 2010-01-12 | 2016-01-12 | International Business Machines Corporation | Customized power rating for computer equipment |
US8352757B2 (en) * | 2010-01-12 | 2013-01-08 | International Business Machines Corporation | Oversubscribing branch circuits while lowering power distribution costs |
JP2011155712A (en) * | 2010-01-25 | 2011-08-11 | Sony Corp | Electronic appliance, power management apparatus, and method of identifying appliance |
US8359598B2 (en) * | 2010-04-01 | 2013-01-22 | Ca, Inc. | Energy efficient scheduling system and method |
US10205319B2 (en) | 2010-06-11 | 2019-02-12 | Eaton Intelligent Power Limited | Automatic matching of sources to loads |
US8805998B2 (en) | 2010-06-11 | 2014-08-12 | Eaton Corporation | Automatic matching of sources to loads |
WO2012025938A1 (en) * | 2010-08-26 | 2012-03-01 | Tata Consultancy Services Limited | Generation of energy consumption profiles |
US20120072752A1 (en) * | 2010-09-22 | 2012-03-22 | Sony Corporation | Method and apparatus for providing power management enhancements |
EP2625585A2 (en) * | 2010-10-04 | 2013-08-14 | Avocent Huntsville Corporation | System and method for monitoring and managing data center resources in real time |
US8909767B2 (en) * | 2010-10-13 | 2014-12-09 | Rackware, Inc. | Cloud federation in a cloud computing environment |
GB2485526A (en) * | 2010-11-04 | 2012-05-23 | Univ Greenwich | Modeling data centre power consumption to determine if a physical server can be virtualized |
US8630822B2 (en) | 2011-02-11 | 2014-01-14 | International Business Machines Corporation | Data center design tool |
US8219258B1 (en) * | 2011-02-25 | 2012-07-10 | eCurv, Inc. | Queuing access to a shared power supply |
US20130325376A1 (en) * | 2011-04-07 | 2013-12-05 | Thomas Edwin Turicchi, Jr. | Systems and methods for determining a power phase and/or a phase rotation |
US8924752B1 (en) | 2011-04-20 | 2014-12-30 | Apple Inc. | Power management for a graphics processing unit or other circuit |
US8589929B2 (en) * | 2011-05-06 | 2013-11-19 | International Business Machines Corporation | System to provide regular and green computing services |
US8589480B2 (en) * | 2011-05-24 | 2013-11-19 | Sony Computer Entertainment America Llc | Automatic performance and capacity measurement for networked servers |
US20130054981A1 (en) | 2011-08-23 | 2013-02-28 | International Business Machines Corporation | Branch Circuit Power Measurement and Dynamic Power Distribution |
US8689220B2 (en) | 2011-11-30 | 2014-04-01 | International Business Machines Corporation | Job scheduling to balance energy consumption and schedule performance |
US8856566B1 (en) | 2011-12-15 | 2014-10-07 | Apple Inc. | Power management scheme that accumulates additional off time for device when no work is available and permits additional power consumption by device when awakened |
US9223630B2 (en) * | 2011-12-22 | 2015-12-29 | Alcatel Lucent | Method and apparatus for energy efficient distributed and elastic load balancing |
US8615763B2 (en) * | 2012-02-01 | 2013-12-24 | Texas Instruments Incorporated | System and method of task allocation in multiprocessing environment based on minimum density or maximum harmonicity that results in lower clock frequency |
US9188109B2 (en) * | 2012-02-16 | 2015-11-17 | Spyros James Lazaris | Virtualization, optimization and adaptation of dynamic demand response in a renewable energy-based electricity grid infrastructure |
US9027141B2 (en) * | 2012-04-12 | 2015-05-05 | Netflix, Inc. | Method and system for improving security and reliability in a networked application environment |
US9390461B1 (en) * | 2012-05-08 | 2016-07-12 | Apple Inc. | Graphics hardware mode controls |
US8843773B2 (en) * | 2012-06-13 | 2014-09-23 | Cisco Technology, Inc. | System and method for automated service profile placement in a network environment |
DE102012017966A1 (en) | 2012-09-12 | 2014-03-13 | Big Dutchman International Gmbh | Method and system for monitoring and / or controlling the resource consumption of an agricultural facility |
US9158345B1 (en) * | 2012-10-15 | 2015-10-13 | Google Inc. | Managing computer performance |
JP6008695B2 (en) * | 2012-11-01 | 2016-10-19 | 株式会社日立製作所 | Power system facility data model conversion method, power system facility data model conversion device, and power system facility data model conversion program |
CN103869923B (en) * | 2012-12-07 | 2018-10-12 | 联想(北京)有限公司 | A kind of method and device of state switching |
US10102594B2 (en) * | 2013-01-09 | 2018-10-16 | International Business Machines Corporation | Dynamic power allocation among points of energy consumption for attaining an objective |
US20140245028A1 (en) * | 2013-02-22 | 2014-08-28 | Qualcomm Incorporated | System and method for temperature driven selection of voltage modes in a portable computing device |
US9594411B2 (en) * | 2013-02-28 | 2017-03-14 | Qualcomm Incorporated | Dynamic power management of context aware services |
US9851726B2 (en) | 2013-09-04 | 2017-12-26 | Panduit Corp. | Thermal capacity management |
US9563724B2 (en) | 2013-09-28 | 2017-02-07 | International Business Machines Corporation | Virtual power management multiprocessor system simulation |
US11749988B2 (en) * | 2014-01-09 | 2023-09-05 | Nautilus True, Llc | System and method for intelligent data center power management and energy market disaster recovery |
KR101828443B1 (en) | 2014-01-28 | 2018-02-13 | 한국전자통신연구원 | User-Friendly Apparatus for Per-Device Power/Energy Measurement and Managemet Method thereof |
US9823673B2 (en) | 2014-04-08 | 2017-11-21 | Qualcomm Incorporated | Energy efficiency aware thermal management in a multi-processor system on a chip based on monitored processing component current draw |
CN105138105A (en) | 2015-07-31 | 2015-12-09 | 北京金山安全软件有限公司 | Power-saving processing method and device, mobile terminal and cloud server |
US20180234491A1 (en) * | 2015-08-14 | 2018-08-16 | Hewlett Packard Enterprise Development Lp | Program deployment according to server efficiency rankings |
US10554519B2 (en) * | 2016-02-08 | 2020-02-04 | Cray Inc. | System and method for dampening power swings in distributed computer environments |
US10430251B2 (en) * | 2016-05-16 | 2019-10-01 | Dell Products L.P. | Systems and methods for load balancing based on thermal parameters |
US10234833B2 (en) * | 2016-07-22 | 2019-03-19 | Intel Corporation | Technologies for predicting power usage of a data center |
US10379893B2 (en) | 2016-08-10 | 2019-08-13 | Rackware, Inc. | Container synchronization |
US11107016B2 (en) * | 2016-08-18 | 2021-08-31 | Virtual Power Systems, Inc. | Augmented power control within a datacenter using predictive modeling |
US20180052574A1 (en) * | 2016-08-22 | 2018-02-22 | United States Of America As Represented By Secretary Of The Navy | Energy Efficiency and Energy Security Optimization Dashboard for Computing Systems |
US11076509B2 (en) | 2017-01-24 | 2021-07-27 | The Research Foundation for the State University | Control systems and prediction methods for it cooling performance in containment |
CN108090563A (en) * | 2017-12-15 | 2018-05-29 | 烟台港股份有限公司 | A kind of electric flux Forecasting Methodology based on BP neural network |
US10739843B2 (en) * | 2018-02-21 | 2020-08-11 | Dell Products L.P. | System and method of monitoring device states |
US10664324B2 (en) | 2018-05-30 | 2020-05-26 | Oracle International Corporation | Intelligent workload migration to optimize power supply efficiencies in computer data centers |
KR102105681B1 (en) * | 2018-10-23 | 2020-04-28 | 울산과학기술원 | Method and apparatus for controlling heterogeneous computing system |
US10922283B2 (en) | 2019-02-22 | 2021-02-16 | Rackware, Inc. | File synchronization |
US11616295B2 (en) | 2019-03-12 | 2023-03-28 | Epirus, Inc. | Systems and methods for adaptive generation of high power electromagnetic radiation and their applications |
US11658410B2 (en) | 2019-03-12 | 2023-05-23 | Epirus, Inc. | Apparatus and method for synchronizing power circuits with coherent RF signals to form a steered composite RF signal |
US11809250B2 (en) * | 2019-10-29 | 2023-11-07 | Intel Corporation | Workload aware power limiting and multiple-input multiple-output control |
KR20210077329A (en) * | 2019-12-17 | 2021-06-25 | 에스케이하이닉스 주식회사 | Storage System, Storage Device and Operating Method Therefor |
US11429176B2 (en) * | 2020-05-14 | 2022-08-30 | Dell Products L.P. | Intelligent and predictive optimization of power needs across virtualized environments |
US11327545B2 (en) * | 2020-06-19 | 2022-05-10 | Dell Products L.P. | Automated management of power distribution during a power crisis |
US20210399700A1 (en) | 2020-06-22 | 2021-12-23 | Epirus, Inc. | Systems and methods for modular power amplifiers |
US20230236656A1 (en) * | 2022-01-26 | 2023-07-27 | Dell Products L.P. | Datacenter carbon footprint climate impact reduction |
CN117032996B (en) * | 2023-10-09 | 2023-12-22 | 湖南中青能科技有限公司 | Power metadata management method and system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080270077A1 (en) * | 2007-04-30 | 2008-10-30 | Mehmet Kivanc Ozonat | System and method for detecting performance anomalies in a computing system |
US20080301474A1 (en) * | 2005-12-23 | 2008-12-04 | Nxp B.V. | Performance Analysis Based System Level Power Management |
US7552350B2 (en) * | 2000-09-27 | 2009-06-23 | Huron Ip Llc | System and method for activity or event base dynamic energy conserving server reconfiguration |
US7802120B2 (en) * | 2005-08-25 | 2010-09-21 | Apple Inc. | Methods and apparatuses for dynamic power control |
US8429431B2 (en) * | 2008-03-07 | 2013-04-23 | Raritan Americas, Inc. | Methods of achieving cognizant power management |
Family Cites Families (131)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4321582A (en) * | 1980-03-11 | 1982-03-23 | Banghart Thomas S | Data retrieval system and method |
US4543649A (en) * | 1983-10-17 | 1985-09-24 | Teknar, Inc. | System for ultrasonically detecting the relative position of a moveable device |
US4955821A (en) * | 1989-07-10 | 1990-09-11 | Litton Systems, Inc. | Method for controlling connector insertion or extraction sequence on power distribution panel |
US5396635A (en) * | 1990-06-01 | 1995-03-07 | Vadem Corporation | Power conservation apparatus having multiple power reduction levels dependent upon the activity of the computer system |
JPH05250188A (en) * | 1992-03-05 | 1993-09-28 | Hitachi Ltd | Priority control system of process |
JPH06221649A (en) | 1993-01-28 | 1994-08-12 | Mitsubishi Electric Corp | Air conditioner |
DE69535577T2 (en) * | 1994-12-22 | 2008-03-06 | Intel Corp., Santa Clara | PERFORMANCE MANAGEMENT WITH CHARACTERISTIC PERFORMANCE CONSUMPTION OF A DEVICE |
US5515853A (en) * | 1995-03-28 | 1996-05-14 | Sonometrics Corporation | Three-dimensional digital ultrasound tracking system |
US5719800A (en) * | 1995-06-30 | 1998-02-17 | Intel Corporation | Performance throttling to reduce IC power consumption |
JPH0996672A (en) | 1995-09-29 | 1997-04-08 | Sukuuea:Kk | Method and system for generating three-dimensional positional data |
US6229899B1 (en) * | 1996-07-17 | 2001-05-08 | American Technology Corporation | Method and device for developing a virtual speaker distant from the sound source |
US7171461B2 (en) * | 1996-07-23 | 2007-01-30 | Server Technology, Inc. | Network remote power management outlet strip |
US7216043B2 (en) * | 1997-02-12 | 2007-05-08 | Power Measurement Ltd. | Push communications architecture for intelligent electronic devices |
AU1249699A (en) | 1997-12-04 | 1999-06-16 | Olivetti Research Limited | Detection system for determining positional information about objects |
US6167330A (en) * | 1998-05-08 | 2000-12-26 | The United States Of America As Represented By The Secretary Of The Air Force | Dynamic power management of systems |
JP3780130B2 (en) | 1998-11-16 | 2006-05-31 | キヤノン株式会社 | Power consumption management device |
GB9901300D0 (en) | 1999-01-22 | 1999-03-10 | Olivetti Research Ltd | A method of increasing the capacity and addressing rate of an Ultrasonic location system |
US6553336B1 (en) * | 1999-06-25 | 2003-04-22 | Telemonitor, Inc. | Smart remote monitoring system and method |
US6499102B1 (en) * | 1999-12-29 | 2002-12-24 | Intel Corporation | Method of dynamically changing the lowest sleeping state in ACPI |
US6413104B1 (en) * | 2000-01-28 | 2002-07-02 | Northrop Grumman Corporation | Power distribution panel with sequence control and enhanced lockout capability |
FR2806473B1 (en) | 2000-03-17 | 2002-09-27 | Pdp Personal Digital Pen | METHODS AND DEVICES FOR ULTRASONIC DETERMINATION OF THE POSITION OF A MOBILE OBJECT |
JP2001337157A (en) | 2000-05-26 | 2001-12-07 | Toyo System Kk | Local positioning system using ultrasonic wave |
US7032119B2 (en) * | 2000-09-27 | 2006-04-18 | Amphus, Inc. | Dynamic power and workload management for multi-server system |
FI111489B (en) * | 2000-12-22 | 2003-07-31 | Iws Int Oy | With intelligent output voltage, intelligent power distribution system works |
DE60209051T2 (en) | 2001-04-20 | 2006-10-26 | Lg Electronics Inc. | Location determination of a mobile communication terminal based on a combination of measurements of signals exchanged between the mobile communication terminal and base stations |
US6567769B2 (en) * | 2001-04-24 | 2003-05-20 | Digipower Manufacturing Inc. | Unattendant data center environment protection, control, and management system |
US7299021B2 (en) * | 2001-12-28 | 2007-11-20 | Nokia Corporation | Method and apparatus for scaling the dynamic range of a receiver for continuously optimizing performance versus power consumption |
US6721672B2 (en) * | 2002-01-02 | 2004-04-13 | American Power Conversion | Method and apparatus for preventing overloads of power distribution networks |
US7606938B2 (en) * | 2002-03-01 | 2009-10-20 | Enterasys Networks, Inc. | Verified device locations in a data network |
JP2003288236A (en) | 2002-03-27 | 2003-10-10 | Nec Software Kyushu Ltd | Management system for large scale integration server |
US20030193777A1 (en) * | 2002-04-16 | 2003-10-16 | Friedrich Richard J. | Data center energy management system |
US6996728B2 (en) * | 2002-04-26 | 2006-02-07 | Hewlett-Packard Development Company, L.P. | Managing power consumption based on utilization statistics |
US6986069B2 (en) * | 2002-07-01 | 2006-01-10 | Newisys, Inc. | Methods and apparatus for static and dynamic power management of computer systems |
US6697300B1 (en) * | 2002-09-13 | 2004-02-24 | General Dynamics Advanced Information Systems, Inc. | Method and apparatus for determining the positioning of volumetric sensor array lines |
US20040064745A1 (en) * | 2002-09-26 | 2004-04-01 | Sudarshan Kadambi | Method and apparatus for controlling the rate at which instructions are executed by a microprocessor system |
NO329096B1 (en) | 2002-12-04 | 2010-08-23 | Sonitor Technologies As | Ultrasonic tracking and location system |
US7210048B2 (en) * | 2003-02-14 | 2007-04-24 | Intel Corporation | Enterprise power and thermal management |
EP1489719A3 (en) | 2003-06-20 | 2007-05-02 | Matsushita Electric Industrial Co., Ltd. | Energy management system, energy management method, and unit for providing information on energy-saving recommended equipment |
US20040267897A1 (en) * | 2003-06-24 | 2004-12-30 | Sychron Inc. | Distributed System Providing Scalable Methodology for Real-Time Control of Server Pools and Data Centers |
JP2005069892A (en) | 2003-08-25 | 2005-03-17 | Toyota Motor Corp | System for computing self-position of moving object |
US6985697B2 (en) * | 2003-09-22 | 2006-01-10 | Nokia, Inc. | Method and system for wirelessly managing the operation of a network appliance over a limited distance |
JP2007510394A (en) * | 2003-10-24 | 2007-04-19 | スクエア・ディー・カンパニー | Intelligent power management control system |
US7134029B2 (en) * | 2003-11-06 | 2006-11-07 | International Business Machines Corporation | Computer-component power-consumption monitoring and control |
MXPA06005159A (en) * | 2003-11-07 | 2007-01-26 | Mpathx Llc | Automatic sensing power systems and methods. |
JP2005198364A (en) | 2003-12-26 | 2005-07-21 | Canon Sales Co Inc | Power supply controller, power supply control method, program for control of power supply, and power supply control system |
US7177728B2 (en) * | 2003-12-30 | 2007-02-13 | Jay Warren Gardner | System and methods for maintaining power usage within a set allocation |
JP2005323438A (en) | 2004-05-07 | 2005-11-17 | Inter Db:Kk | Power control system |
US7418608B2 (en) * | 2004-06-17 | 2008-08-26 | Intel Corporation | Method and an apparatus for managing power consumption of a server |
US7281146B2 (en) * | 2004-06-30 | 2007-10-09 | Intel Corporation | Dynamic power requirement budget manager |
JP2006025474A (en) | 2004-07-06 | 2006-01-26 | Chugoku Electric Power Co Inc:The | Method for managing power supply to electric apparatus and power management network system |
ATE433611T1 (en) * | 2004-10-05 | 2009-06-15 | 2D2C Inc | ELECTRICAL POWER DISTRIBUTION SYSTEM |
US8108929B2 (en) * | 2004-10-19 | 2012-01-31 | Reflex Systems, LLC | Method and system for detecting intrusive anomalous use of a software system using multiple detection algorithms |
US7333000B2 (en) * | 2004-11-12 | 2008-02-19 | Afco Systems Development, Inc. | Tracking system and method for electrically powered equipment |
US20060112286A1 (en) * | 2004-11-23 | 2006-05-25 | Whalley Ian N | Method for dynamically reprovisioning applications and other server resources in a computer center in response to power and heat dissipation requirements |
US9182480B2 (en) | 2005-01-28 | 2015-11-10 | Hewlett-Packard Development Company, L.P. | Information technology (IT) equipment positioning system |
US20060168975A1 (en) * | 2005-01-28 | 2006-08-03 | Hewlett-Packard Development Company, L.P. | Thermal and power management apparatus |
US20060184936A1 (en) * | 2005-02-11 | 2006-08-17 | Timothy Abels | System and method using virtual machines for decoupling software from management and control systems |
US20060184937A1 (en) | 2005-02-11 | 2006-08-17 | Timothy Abels | System and method for centralized software management in virtual machines |
US20060184935A1 (en) * | 2005-02-11 | 2006-08-17 | Timothy Abels | System and method using virtual machines for decoupling software from users and services |
FR2882596B1 (en) | 2005-02-25 | 2009-07-10 | Eads Ccr Groupement D Interet | LOCATION OF A NON-DESTRUCTIVE CONTROL PROBE |
US7881910B2 (en) * | 2005-05-02 | 2011-02-01 | American Power Conversion Corporation | Methods and systems for managing facility power and cooling |
US7885795B2 (en) * | 2005-05-02 | 2011-02-08 | American Power Conversion Corporation | Methods and systems for managing facility power and cooling |
WO2006119248A2 (en) | 2005-05-02 | 2006-11-09 | American Power Conversion Corporation | Methods and systems for managing facility power and cooling |
US7895012B2 (en) * | 2005-05-03 | 2011-02-22 | Hewlett-Packard Development Company, L.P. | Systems and methods for organizing and storing data |
US7644148B2 (en) | 2005-05-16 | 2010-01-05 | Hewlett-Packard Development Company, L.P. | Historical data based workload allocation |
US7444526B2 (en) * | 2005-06-16 | 2008-10-28 | International Business Machines Corporation | Performance conserving method for reducing power consumption in a server system |
US7613807B2 (en) * | 2005-07-04 | 2009-11-03 | Hewlett-Packard Development Company, L.P. | System, method, and apparatus for discovering a new server connected within an automated data center |
US7869965B2 (en) | 2005-08-17 | 2011-01-11 | Oracle America, Inc. | Inferential power monitor without voltage/current transducers |
US8374730B2 (en) * | 2005-08-25 | 2013-02-12 | Apple Inc. | Methods and apparatuses for dynamic thermal control |
JP2007139523A (en) | 2005-11-16 | 2007-06-07 | Ntt Facilities Inc | Power consumption presentation system, server device, and power consumption presentation method |
JP4895266B2 (en) * | 2005-12-28 | 2012-03-14 | 富士通株式会社 | Management system, management program, and management method |
US8494661B2 (en) * | 2007-12-28 | 2013-07-23 | Server Technology, Inc. | Power distribution, management, and monitoring systems and methods |
US7561411B2 (en) | 2006-03-17 | 2009-07-14 | Eaton Corporation | Uninterruptible power distribution systems and methods using distributed power distribution units |
JP2007299624A (en) | 2006-04-28 | 2007-11-15 | Dainippon Printing Co Ltd | Receptacle system with display device |
US7669071B2 (en) * | 2006-05-05 | 2010-02-23 | Dell Products L.P. | Power allocation management in an information handling system |
JP2007330016A (en) | 2006-06-07 | 2007-12-20 | Nec Corp | Power unit, information processing system, and method and program for controlling power unit |
US20080052145A1 (en) * | 2006-08-10 | 2008-02-28 | V2 Green, Inc. | Power Aggregation System for Distributed Electric Resources |
JPWO2008026463A1 (en) | 2006-08-30 | 2010-01-21 | 日本電気株式会社 | Position measurement system, robot, position measurement method, and sound source position calculation program |
US20080148075A1 (en) * | 2006-12-15 | 2008-06-19 | Texas Instruments Incorporated | Method and system of controlling power states of devices |
US7969156B2 (en) * | 2007-03-30 | 2011-06-28 | Liebert Corporation | Method and apparatus for monitoring a load |
US7783910B2 (en) * | 2007-03-30 | 2010-08-24 | International Business Machines Corporation | Method and system for associating power consumption of a server with a network address assigned to the server |
TWI327230B (en) | 2007-04-03 | 2010-07-11 | Ind Tech Res Inst | Sound source localization system and sound soure localization method |
US8880907B2 (en) * | 2007-06-21 | 2014-11-04 | Schneider Electric It Corporation | Method and system for determining physical location of equipment |
US7796471B2 (en) * | 2008-02-20 | 2010-09-14 | Intelligent Sciences, Ltd. | Ultrasonic in-building positioning system based on phase difference array with ranging |
CA2740781C (en) * | 2008-10-20 | 2016-06-14 | Raritan Americas, Inc. | System and method for automatic determination of the physical location of data center equipment |
US20120133510A1 (en) | 2010-11-30 | 2012-05-31 | Panduit Corp. | Physical infrastructure management system having an integrated cabinet |
US9715222B2 (en) | 2011-02-09 | 2017-07-25 | Avocent Huntsville, Llc | Infrastructure control fabric system and method |
US9746911B2 (en) | 2011-03-02 | 2017-08-29 | Tso Logic Inc. | Same linking |
US9185824B2 (en) | 2011-04-26 | 2015-11-10 | Panduit Corp. | Cable pathway system for network architecture |
US8924548B2 (en) | 2011-08-16 | 2014-12-30 | Panduit Corp. | Integrated asset tracking, task manager, and virtual container for data center management |
CN204613978U (en) | 2011-10-03 | 2015-09-02 | 阿沃森特亨茨维尔公司 | For managing the system of the assets in specific environment |
US9661016B2 (en) | 2011-12-06 | 2017-05-23 | Avocent Huntsville Corp. | Data center infrastructure management system incorporating security for managed infrastructure devices |
EP2791825B1 (en) | 2011-12-12 | 2017-06-21 | Avocent Huntsville, LLC | System and method for monitoring and managing data center resources in real time incorporating manageability subsystem |
US8938634B2 (en) | 2012-01-25 | 2015-01-20 | Empire Technology Development Llc | User generated data center power savings |
KR101586469B1 (en) | 2012-02-01 | 2016-02-02 | 엠파이어 테크놀로지 디벨롭먼트 엘엘씨 | Preventing cloud cartography |
US9463574B2 (en) | 2012-03-01 | 2016-10-11 | Irobot Corporation | Mobile inspection robot |
WO2013147812A1 (en) | 2012-03-29 | 2013-10-03 | Empire Technology Development, Llc | Resource management for data center based gaming |
US9183019B2 (en) | 2012-04-25 | 2015-11-10 | Empire Technology Development Llc | Certification for flexible resource demand applications |
US9537879B2 (en) | 2012-06-21 | 2017-01-03 | Avocent Huntsville, Llc | Cyber security monitoring system and method for data center components |
EP2688003A1 (en) | 2012-07-16 | 2014-01-22 | Fujitsu Limited | Method, apparatus, and program for generating a simulation model of a space |
US8917513B1 (en) | 2012-07-30 | 2014-12-23 | Methode Electronics, Inc. | Data center equipment cabinet information center and updateable asset tracking system |
US8917512B2 (en) | 2012-07-30 | 2014-12-23 | Methode Electronics, Inc. | Data center equipment cabinet information center |
US8610438B1 (en) | 2012-08-09 | 2013-12-17 | Precision Air & Energy Services, LLC | Branch circuit monitor |
CN104704911B (en) | 2012-10-11 | 2019-12-24 | 维谛信息技术系统有限公司 | System and method for tracking assets in conjunction with a wireless network |
US10693052B2 (en) | 2012-11-26 | 2020-06-23 | Abb Schweiz Ag | System and method for energy harvesting in a data center |
KR101981258B1 (en) | 2013-01-04 | 2019-05-22 | 삼성전자주식회사 | Method for sharing contents using personal cloud device, Electronic device and Personal Cloud System thereof |
US8964360B2 (en) | 2013-02-06 | 2015-02-24 | Jonathan D. Trout | System to connect and multiplex sensor signals |
KR102056128B1 (en) | 2013-02-28 | 2019-12-17 | 삼성전자 주식회사 | Portable apparatus and method for taking a photograph by using widget |
US9454501B2 (en) | 2013-07-30 | 2016-09-27 | Leviton Manufacturing Co., Inc. | Intelligent patch panel port monitoring system |
US9784460B2 (en) | 2013-08-01 | 2017-10-10 | Nautilus Data Technologies, Inc. | Data center facility and process that utilizes a closed-looped heat management system |
US9851726B2 (en) | 2013-09-04 | 2017-12-26 | Panduit Corp. | Thermal capacity management |
US9218662B1 (en) | 2014-11-06 | 2015-12-22 | Duelight Llc | System, method, and computer program product for exchanging images |
US9460118B2 (en) | 2014-09-30 | 2016-10-04 | Duelight Llc | System, method, and computer program product for exchanging images |
GB2519547A (en) | 2013-10-24 | 2015-04-29 | Eaton Ind France Sas | Method of controlling a data centre architecture equipment |
JP6221649B2 (en) | 2013-11-06 | 2017-11-01 | 東ソー株式会社 | Temperature controllable analyzer |
US20150178401A1 (en) | 2013-12-19 | 2015-06-25 | Panduit Corp. | Methods and systems for identifying stranded capacity |
US9367422B2 (en) | 2014-01-15 | 2016-06-14 | Ca, Inc. | Determining and using power utilization indexes for servers |
US20150212126A1 (en) | 2014-01-26 | 2015-07-30 | Abb Technology Ag | System and method for unified power quality monitoring and data collection in a power system having heterogeneous devices for monitoring power quality |
US9727515B2 (en) | 2014-02-19 | 2017-08-08 | Cyber Switching Patents, Llc | Cabinet level controller with asset management |
US9632835B2 (en) | 2014-03-17 | 2017-04-25 | Ca, Inc. | Deployment of virtual machines to physical host machines based on infrastructure utilization decisions |
US9448841B2 (en) | 2014-03-17 | 2016-09-20 | Ca, Inc. | Decommissioning physical host machines after relocating unproductive virtual machines therefrom |
US9405569B2 (en) | 2014-03-17 | 2016-08-02 | Ca, Inc. | Determining virtual machine utilization of distributed computed system infrastructure |
US9448848B2 (en) | 2014-03-17 | 2016-09-20 | Ca, Inc. | Controlling placement of virtual machines on physical host machines and placement of physical host machines in cabinets |
US9461709B2 (en) | 2014-06-30 | 2016-10-04 | Intel Corporation | Methods and systems for server power line communication |
US9943011B2 (en) | 2014-09-10 | 2018-04-10 | Panduit Corp. | Cooling control for data centers with cold aisle containment systems |
US9769254B2 (en) | 2014-09-15 | 2017-09-19 | Ca, Inc. | Productive spend metric based resource management for a portfolio of distributed computing systems |
CA2911343A1 (en) | 2014-11-06 | 2016-05-06 | Cinnos Technologies, Inc. | Smart mission critical rack |
US9870773B2 (en) | 2014-11-11 | 2018-01-16 | Commscope, Inc. Of North Carolina | Multi-tier intelligent infrastructure management systems for communications systems and related equipment and methods |
AU2016258540A1 (en) | 2015-05-01 | 2017-11-02 | Baselayer Technology, Llc | Side-cooled modular data center |
GB201513039D0 (en) | 2015-07-23 | 2015-09-09 | Eaton Ind France Sas | Shutting down of a virtual system |
US9568923B1 (en) | 2015-10-27 | 2017-02-14 | International Business Machines Corporation | Determining a time for corrective action in a data center |
-
2009
- 2009-10-21 US US12/603,322 patent/US8429431B2/en active Active
- 2009-10-21 AU AU2009308467A patent/AU2009308467A1/en not_active Abandoned
- 2009-10-21 JP JP2011533304A patent/JP2012506597A/en active Pending
- 2009-10-21 EP EP09822648A patent/EP2350770A4/en not_active Withdrawn
- 2009-10-21 CN CN2009801490753A patent/CN102227693A/en active Pending
- 2009-10-21 CA CA2741088A patent/CA2741088C/en active Active
- 2009-10-21 WO PCT/US2009/061521 patent/WO2010048316A1/en active Application Filing
-
2013
- 2013-03-08 US US13/790,601 patent/US20130191658A1/en not_active Abandoned
-
2015
- 2015-10-07 US US14/877,621 patent/US10289184B2/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7552350B2 (en) * | 2000-09-27 | 2009-06-23 | Huron Ip Llc | System and method for activity or event base dynamic energy conserving server reconfiguration |
US7802120B2 (en) * | 2005-08-25 | 2010-09-21 | Apple Inc. | Methods and apparatuses for dynamic power control |
US20080301474A1 (en) * | 2005-12-23 | 2008-12-04 | Nxp B.V. | Performance Analysis Based System Level Power Management |
US20080270077A1 (en) * | 2007-04-30 | 2008-10-30 | Mehmet Kivanc Ozonat | System and method for detecting performance anomalies in a computing system |
US8429431B2 (en) * | 2008-03-07 | 2013-04-23 | Raritan Americas, Inc. | Methods of achieving cognizant power management |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8811377B1 (en) | 2010-08-30 | 2014-08-19 | Synapsense Corporation | Apparatus and method for instrumenting devices to measure power usage using a multi-tier wireless network |
US20120078429A1 (en) * | 2010-09-24 | 2012-03-29 | Patrick Edward Weston | APPARATUS AND METHOD FOR COLLECTING AND DISTRIBUTING POWER USAGE DATA FROM RACK POWER DISTRIBUTION UNITS (RPDUs) USING A WIRELESS SENSOR NETWORK |
US8600575B2 (en) * | 2010-09-24 | 2013-12-03 | Synapsense Corporation | Apparatus and method for collecting and distributing power usage data from rack power distribution units (RPDUs) using a wireless sensor network |
US20130151696A1 (en) * | 2011-12-12 | 2013-06-13 | Delta Electronics, Inc. | Trigger method of computational procedure for virtual maching migration and application program for the same |
US8903992B2 (en) * | 2011-12-12 | 2014-12-02 | Delta Electronics, Inc. | Trigger method of computational procedure for virtual machine migration and application program for the same |
US9154367B1 (en) * | 2011-12-27 | 2015-10-06 | Google Inc. | Load balancing and content preservation |
US20140365641A1 (en) * | 2013-06-11 | 2014-12-11 | Samsung Electronics Co., Ltd | Processor module, server system and method of controlling processor module |
WO2016021986A1 (en) * | 2014-08-07 | 2016-02-11 | Samsung Electronics Co., Ltd. | Method and apparatus for executing software in electronic device |
US9904582B2 (en) | 2014-08-07 | 2018-02-27 | Samsung Electronics Co., Ltd. | Method and apparatus for executing software in electronic device |
US9696781B2 (en) * | 2015-05-28 | 2017-07-04 | Cisco Technology, Inc. | Automated power control for reducing power usage in communications networks |
WO2020246985A1 (en) * | 2019-06-07 | 2020-12-10 | Hewlett-Packard Development Company, L.P. | Device operation modifications based on monitored power levels |
US11269677B2 (en) | 2019-08-08 | 2022-03-08 | Vigyanlabs Innovations Private Limited | System and method to analyze and optimize application level resource and energy consumption by data center servers |
Also Published As
Publication number | Publication date |
---|---|
US8429431B2 (en) | 2013-04-23 |
JP2012506597A (en) | 2012-03-15 |
US10289184B2 (en) | 2019-05-14 |
CA2741088A1 (en) | 2010-04-29 |
EP2350770A4 (en) | 2012-09-05 |
CN102227693A (en) | 2011-10-26 |
US20100235654A1 (en) | 2010-09-16 |
US20160252946A1 (en) | 2016-09-01 |
AU2009308467A1 (en) | 2010-04-29 |
WO2010048316A1 (en) | 2010-04-29 |
EP2350770A1 (en) | 2011-08-03 |
CA2741088C (en) | 2017-07-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10289184B2 (en) | Methods of achieving cognizant power management | |
CA2722946C (en) | Environmentally cognizant power management | |
Vasan et al. | Worth their watts?-an empirical study of datacenter servers | |
Guenter et al. | Managing cost, performance, and reliability tradeoffs for energy-aware server provisioning | |
Cai et al. | SLA-aware energy-efficient scheduling scheme for Hadoop YARN | |
Gu et al. | Energy efficient scheduling of servers with multi-sleep modes for cloud data center | |
Cai et al. | Energy-aware high performance computing: A taxonomy study | |
Moreno et al. | Customer-aware resource overallocation to improve energy efficiency in realtime cloud computing data centers | |
Bartolini et al. | Paving the way toward energy-aware and automated datacentre | |
Kumar et al. | Heterogeneity and thermal aware adaptive heuristics for energy efficient consolidation of virtual machines in infrastructure clouds | |
Jin et al. | Energy-efficient task scheduling for CPU-intensive streaming jobs on Hadoop | |
Lent | Analysis of an energy proportional data center | |
Hou et al. | Power grab in aggressively provisioned data centers: What is the risk and what can be done about it | |
de Assunçao et al. | An analysis of power consumption logs from a monitored grid site | |
Kalogirou et al. | Exploiting CPU voltage margins to increase the profit of cloud infrastructure providers | |
Karpowicz et al. | Energy and power efficiency in cloud | |
Mehrotra et al. | Power-Aware Modeling and Autonomic Management Framework for Distributed Computing Systems. | |
Mehrotra et al. | A power-aware autonomic approach for performance management of scientific applications in a data center environment | |
Chen et al. | A cyber-physical integrated system for application performance and energy management in data centers | |
Krzywda et al. | Modeling and simulation of qos-aware power budgeting in cloud data centers | |
Singh et al. | Energy aware VM consolidation using dynamic threshold in cloud computing | |
Bolla et al. | Enhancing energy-efficient cloud management through code annotations and the green abstraction layer | |
Islam et al. | Distributed resource management in data center with temperature constraint | |
Bhattacharya | Constraints And Techniques For Software Power Management In Production Clusters | |
Jiang et al. | Dynamic voltage/frequency scaling for power reduction in data centers: Enough or not? |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: RARITAN AMERICAS, INC., NEW JERSEY Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:MALIK, NAIM R.;PAETZ, CHRISTIAN, DR.;WEINSTOCK, NEIL;AND OTHERS;SIGNING DATES FROM 20101230 TO 20110815;REEL/FRAME:033006/0069 |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO PAY ISSUE FEE |